Financial Analytics of Stock Returns during Covid-19 Pandemic in 2020.ipynb

{"nbformat":4,"nbformat_minor":0,"metadata":{"colab":{"name":"Individual Project- Tzu-Hua Huang.ipynb","provenance":[],"collapsed_sections":["RwKMHPWNdQnV"],"toc_visible":true,"authorship_tag":"ABX9TyM69gOt72rDGEti1F6KtLqw"},"kernelspec":{"name":"python3","display_name":"Python 3"},"language_info":{"name":"python"}},"cells":[{"cell_type":"markdown","metadata":{"id":"uzkdsgH0_zHS"},"source":["Name: Tzu-Hua Huang"]},{"cell_type":"markdown","metadata":{"id":"Ky5ngGo-heXK"},"source":["## Set up"]},{"cell_type":"code","metadata":{"colab":{"base_uri":"https://localhost:8080/"},"id":"t8d2MZLFFXV8","executionInfo":{"status":"ok","timestamp":1625713106175,"user_tz":240,"elapsed":194,"user":{"displayName":"Tzu-Hua Huang","photoUrl":"https://lh3.googleusercontent.com/a-/AOh14GgbXOdZWwXGVTDuMHbeSme5C5oKAeF6I6ojY6qc=s64","userId":"01458545184256376706"}},"outputId":"bf52cd92-e3be-4e81-c0a2-bda66c694d53"},"source":["# Mount Google Drive\n","from google.colab import drive\n","drive.mount('/content/gdrive')"],"execution_count":161,"outputs":[{"output_type":"stream","text":["Drive already mounted at /content/gdrive; to attempt to forcibly remount, call drive.mount(\"/content/gdrive\", force_remount=True).\n"],"name":"stdout"}]},{"cell_type":"code","metadata":{"colab":{"base_uri":"https://localhost:8080/"},"id":"K2-wwtXMHF2v","executionInfo":{"status":"ok","timestamp":1625713106651,"user_tz":240,"elapsed":260,"user":{"displayName":"Tzu-Hua Huang","photoUrl":"https://lh3.googleusercontent.com/a-/AOh14GgbXOdZWwXGVTDuMHbeSme5C5oKAeF6I6ojY6qc=s64","userId":"01458545184256376706"}},"outputId":"95a07912-54a9-4b8a-fc1c-54348ac53b7e"},"source":["import os\n","\n","# the base Google Drive directory\n","root_dir = \"/content/gdrive/MyDrive/Colab Notebooks/BA870/\"\n","\n","# choose where you want your project files to be saved\n","project_folder = \"Individual Project/\"\n","\n","def create_and_set_working_directory(project_folder):\n","  # check if your project folder exists. if not, it will be created.\n","  if os.path.isdir(root_dir + project_folder) == False:\n","    os.mkdir(root_dir + project_folder)\n","    print(root_dir + project_folder + ' did not exist but was created.')\n","\n","  # change the OS to use your project folder as the working directory\n","  os.chdir(root_dir + project_folder)\n","\n","  # create a test file to make sure it shows up in the right place\n","  !touch 'new_file_in_working_directory.txt'\n","  print('\\nYour working directory was changed to ' + root_dir + project_folder + \\\n","        \"\\n\\nAn empty text file was created there. You can also run !pwd to confirm the current working directory.\" )\n","\n","create_and_set_working_directory(project_folder)"],"execution_count":162,"outputs":[{"output_type":"stream","text":["\n","Your working directory was changed to /content/gdrive/MyDrive/Colab Notebooks/BA870/Individual Project/\n","\n","An empty text file was created there. You can also run !pwd to confirm the current working directory.\n"],"name":"stdout"}]},{"cell_type":"code","metadata":{"colab":{"base_uri":"https://localhost:8080/"},"id":"cMqJX54hze7i","executionInfo":{"status":"ok","timestamp":1625713106849,"user_tz":240,"elapsed":200,"user":{"displayName":"Tzu-Hua Huang","photoUrl":"https://lh3.googleusercontent.com/a-/AOh14GgbXOdZWwXGVTDuMHbeSme5C5oKAeF6I6ojY6qc=s64","userId":"01458545184256376706"}},"outputId":"322d0d6e-8850-46ca-990d-e834d6aac62f"},"source":["!pwd"],"execution_count":163,"outputs":[{"output_type":"stream","text":["/content/gdrive/MyDrive/Colab Notebooks/BA870/Individual Project\n"],"name":"stdout"}]},{"cell_type":"markdown","metadata":{"id":"hd5I0KZR6iOu"},"source":["Check if the function worked by listing the files in the project folder. During the very first run it should contain only `new_file_in_working_directory.txt`. If you upload other files to your `project_folder`, other files will be listed too. "]},{"cell_type":"code","metadata":{"colab":{"base_uri":"https://localhost:8080/"},"id":"kXsOSNC96kRp","executionInfo":{"status":"ok","timestamp":1625713107067,"user_tz":240,"elapsed":227,"user":{"displayName":"Tzu-Hua Huang","photoUrl":"https://lh3.googleusercontent.com/a-/AOh14GgbXOdZWwXGVTDuMHbeSme5C5oKAeF6I6ojY6qc=s64","userId":"01458545184256376706"}},"outputId":"0b03f0a2-c297-45df-80e2-36fc43c3fff6"},"source":["!ls"],"execution_count":164,"outputs":[{"output_type":"stream","text":["'Individual Project- Tzu-Hua Huang.ipynb'   tempGICS.csv\n"," new_file_in_working_directory.txt\t    tempNAICS.csv\n"," Return2020.csv\t\t\t\t    tempSIC.csv\n"," Return2020_w_Industry_Codes.csv\t    WRDS_Variables.csv\n"],"name":"stdout"}]},{"cell_type":"markdown","metadata":{"id":"s95UImxR8ba0"},"source":["Please upload the necessary files (`stock_des.csv`, `ta.csv`, `rev.csv`) in your `project_folder` folder. \n","\n","You can upload the files in your Google Drive `project_folder` via drive.google.com or via below `files.upload()` button. "]},{"cell_type":"code","metadata":{"id":"LZ71uhIpKDI_","executionInfo":{"status":"ok","timestamp":1625713107085,"user_tz":240,"elapsed":84,"user":{"displayName":"Tzu-Hua Huang","photoUrl":"https://lh3.googleusercontent.com/a-/AOh14GgbXOdZWwXGVTDuMHbeSme5C5oKAeF6I6ojY6qc=s64","userId":"01458545184256376706"}}},"source":["# imports\n","import pandas as pd\n","import numpy as np\n","import statsmodels.api as sm\n","import matplotlib.pyplot as plt\n","import seaborn as sns"],"execution_count":165,"outputs":[]},{"cell_type":"markdown","metadata":{"id":"jaZYJ63fzJf0"},"source":["# 1.2019 Financial Accounting Characteristics that Explain Varying Exposures to Covid-19 Market Shock (for Early and Late 2020\n"]},{"cell_type":"markdown","metadata":{"id":"3LBtv8DDfik2"},"source":["## Step 1. Load and merge data\n","\n","Load the data from `temp1.csv` and `temp2.csv` into Pandas dataframes. Merge the datasets based on each stock's ticker symbol (TIC)."]},{"cell_type":"code","metadata":{"colab":{"base_uri":"https://localhost:8080/","height":204},"id":"TlfPhFQy6xCt","executionInfo":{"status":"ok","timestamp":1625713107088,"user_tz":240,"elapsed":81,"user":{"displayName":"Tzu-Hua Huang","photoUrl":"https://lh3.googleusercontent.com/a-/AOh14GgbXOdZWwXGVTDuMHbeSme5C5oKAeF6I6ojY6qc=s64","userId":"01458545184256376706"}},"outputId":"aad837f7-bc1a-4298-cd0f-7bd287ad7b19"},"source":["# load data including RetEarly2020 and RetLate2020\n","temp1 = pd.read_csv('Return2020.csv')\n","temp1.head()"],"execution_count":166,"outputs":[{"output_type":"execute_result","data":{"text/html":["<div>\n","<style scoped>\n","    .dataframe tbody tr th:only-of-type {\n","        vertical-align: middle;\n","    }\n","\n","    .dataframe tbody tr th {\n","        vertical-align: top;\n","    }\n","\n","    .dataframe thead th {\n","        text-align: right;\n","    }\n","</style>\n","<table border=\"1\" class=\"dataframe\">\n","  <thead>\n","    <tr style=\"text-align: right;\">\n","      <th></th>\n","      <th>TICKER</th>\n","      <th>RetEarly2020</th>\n","      <th>RetLate2020</th>\n","    </tr>\n","  </thead>\n","  <tbody>\n","    <tr>\n","      <th>0</th>\n","      <td>JJSF</td>\n","      <td>-0.340234493</td>\n","      <td>0.300340</td>\n","    </tr>\n","    <tr>\n","      <th>1</th>\n","      <td>PLXS</td>\n","      <td>-0.290875889</td>\n","      <td>0.433469</td>\n","    </tr>\n","    <tr>\n","      <th>2</th>\n","      <td>HNGR</td>\n","      <td>-0.435711705</td>\n","      <td>0.411425</td>\n","    </tr>\n","    <tr>\n","      <th>3</th>\n","      <td>ORCL</td>\n","      <td>-0.08359557</td>\n","      <td>0.356166</td>\n","    </tr>\n","    <tr>\n","      <th>4</th>\n","      <td>MSFT</td>\n","      <td>0.003212116</td>\n","      <td>0.421139</td>\n","    </tr>\n","  </tbody>\n","</table>\n","</div>"],"text/plain":["  TICKER  RetEarly2020  RetLate2020\n","0   JJSF  -0.340234493     0.300340\n","1   PLXS  -0.290875889     0.433469\n","2   HNGR  -0.435711705     0.411425\n","3   ORCL   -0.08359557     0.356166\n","4   MSFT   0.003212116     0.421139"]},"metadata":{"tags":[]},"execution_count":166}]},{"cell_type":"code","metadata":{"colab":{"base_uri":"https://localhost:8080/","height":219},"id":"WIp2_wIxryfg","executionInfo":{"status":"ok","timestamp":1625713107090,"user_tz":240,"elapsed":80,"user":{"displayName":"Tzu-Hua Huang","photoUrl":"https://lh3.googleusercontent.com/a-/AOh14GgbXOdZWwXGVTDuMHbeSme5C5oKAeF6I6ojY6qc=s64","userId":"01458545184256376706"}},"outputId":"b712d15a-11f1-404c-c1ec-c09648b00480"},"source":["# check data\n","temp1.info()\n","temp1[temp1.isnull().any(axis=1)]\n","# temp1['RetEarly2020'] = pd.to_numeric(temp1['RetEarly2020']) # -- ValueError: Unable to parse string \"#VALUE!\""],"execution_count":167,"outputs":[{"output_type":"stream","text":["<class 'pandas.core.frame.DataFrame'>\n","RangeIndex: 2716 entries, 0 to 2715\n","Data columns (total 3 columns):\n"," #   Column        Non-Null Count  Dtype  \n","---  ------        --------------  -----  \n"," 0   TICKER        2716 non-null   object \n"," 1   RetEarly2020  2716 non-null   object \n"," 2   RetLate2020   2716 non-null   float64\n","dtypes: float64(1), object(2)\n","memory usage: 63.8+ KB\n"],"name":"stdout"},{"output_type":"execute_result","data":{"text/html":["<div>\n","<style scoped>\n","    .dataframe tbody tr th:only-of-type {\n","        vertical-align: middle;\n","    }\n","\n","    .dataframe tbody tr th {\n","        vertical-align: top;\n","    }\n","\n","    .dataframe thead th {\n","        text-align: right;\n","    }\n","</style>\n","<table border=\"1\" class=\"dataframe\">\n","  <thead>\n","    <tr style=\"text-align: right;\">\n","      <th></th>\n","      <th>TICKER</th>\n","      <th>RetEarly2020</th>\n","      <th>RetLate2020</th>\n","    </tr>\n","  </thead>\n","  <tbody>\n","  </tbody>\n","</table>\n","</div>"],"text/plain":["Empty DataFrame\n","Columns: [TICKER, RetEarly2020, RetLate2020]\n","Index: []"]},"metadata":{"tags":[]},"execution_count":167}]},{"cell_type":"markdown","metadata":{"id":"5s7xWGkxr9HC"},"source":["`ValueError` occurs when converting the datatype of `RetEarly2020`. Thus, we drop the rows with errors accroding to the warning message."]},{"cell_type":"code","metadata":{"colab":{"base_uri":"https://localhost:8080/"},"id":"sQ0dMsHssIdG","executionInfo":{"status":"ok","timestamp":1625713107094,"user_tz":240,"elapsed":65,"user":{"displayName":"Tzu-Hua Huang","photoUrl":"https://lh3.googleusercontent.com/a-/AOh14GgbXOdZWwXGVTDuMHbeSme5C5oKAeF6I6ojY6qc=s64","userId":"01458545184256376706"}},"outputId":"6e46baf2-a526-455a-d4a9-3710ad55fa12"},"source":["# drop the rows with errors\n","temp1.drop(temp1.loc[temp1['RetEarly2020'] == '#VALUE!'].index, inplace=True)\n","\n","# convert datatype of RetEarly2020 from object to float\n","temp1['RetEarly2020'] = pd.to_numeric(temp1['RetEarly2020'])\n","temp1.info()"],"execution_count":168,"outputs":[{"output_type":"stream","text":["<class 'pandas.core.frame.DataFrame'>\n","Int64Index: 2711 entries, 0 to 2715\n","Data columns (total 3 columns):\n"," #   Column        Non-Null Count  Dtype  \n","---  ------        --------------  -----  \n"," 0   TICKER        2711 non-null   object \n"," 1   RetEarly2020  2711 non-null   float64\n"," 2   RetLate2020   2711 non-null   float64\n","dtypes: float64(2), object(1)\n","memory usage: 84.7+ KB\n"],"name":"stdout"}]},{"cell_type":"markdown","metadata":{"id":"jrDPSKM_sPLO"},"source":["Download required variables for creating financial ratios from WRDS Compustat. Load data set `WRDS_Variables.csv` and save as `temp2`.\n"]},{"cell_type":"code","metadata":{"colab":{"base_uri":"https://localhost:8080/","height":723},"id":"8S0i8l1DxBrz","executionInfo":{"status":"ok","timestamp":1625713107411,"user_tz":240,"elapsed":369,"user":{"displayName":"Tzu-Hua Huang","photoUrl":"https://lh3.googleusercontent.com/a-/AOh14GgbXOdZWwXGVTDuMHbeSme5C5oKAeF6I6ojY6qc=s64","userId":"01458545184256376706"}},"outputId":"e26fba06-2b24-4509-e582-56e05b22fe40"},"source":["# load data #2\n","temp2 = pd.read_csv('WRDS_Variables.csv')\n","\n","# display all rows of dataframe\n","# pd.set_option('display.max_rows', None)\n","temp2.sample(20)"],"execution_count":169,"outputs":[{"output_type":"execute_result","data":{"text/html":["<div>\n","<style scoped>\n","    .dataframe tbody tr th:only-of-type {\n","        vertical-align: middle;\n","    }\n","\n","    .dataframe tbody tr th {\n","        vertical-align: top;\n","    }\n","\n","    .dataframe thead th {\n","        text-align: right;\n","    }\n","</style>\n","<table border=\"1\" class=\"dataframe\">\n","  <thead>\n","    <tr style=\"text-align: right;\">\n","      <th></th>\n","      <th>gvkey</th>\n","      <th>datadate</th>\n","      <th>fyear</th>\n","      <th>indfmt</th>\n","      <th>consol</th>\n","      <th>popsrc</th>\n","      <th>datafmt</th>\n","      <th>tic</th>\n","      <th>curcd</th>\n","      <th>ogm</th>\n","      <th>act</th>\n","      <th>at</th>\n","      <th>che</th>\n","      <th>dltt</th>\n","      <th>ebit</th>\n","      <th>emp</th>\n","      <th>intan</th>\n","      <th>invt</th>\n","      <th>lct</th>\n","      <th>lt</th>\n","      <th>ni</th>\n","      <th>ppent</th>\n","      <th>re</th>\n","      <th>revt</th>\n","      <th>sale</th>\n","      <th>xint</th>\n","      <th>xintd</th>\n","      <th>costat</th>\n","      <th>city</th>\n","      <th>ggroup</th>\n","      <th>loc</th>\n","      <th>state</th>\n","      <th>ipodate</th>\n","    </tr>\n","  </thead>\n","  <tbody>\n","    <tr>\n","      <th>1392</th>\n","      <td>28742</td>\n","      <td>20191231</td>\n","      <td>2019.0</td>\n","      <td>INDL</td>\n","      <td>C</td>\n","      <td>D</td>\n","      <td>STD</td>\n","      <td>BWA</td>\n","      <td>USD</td>\n","      <td>NaN</td>\n","      <td>3836.000</td>\n","      <td>9702.000</td>\n","      <td>832.000</td>\n","      <td>1741.000</td>\n","      <td>1228.000</td>\n","      <td>29.000</td>\n","      <td>2244.000</td>\n","      <td>807.000</td>\n","      <td>2329.000</td>\n","      <td>4858.000</td>\n","      <td>746.000</td>\n","      <td>3010.000</td>\n","      <td>5215.000</td>\n","      <td>10168.000</td>\n","      <td>10168.000</td>\n","      <td>71.000</td>\n","      <td>NaN</td>\n","      <td>A</td>\n","      <td>Auburn Hills</td>\n","      <td>2510</td>\n","      <td>USA</td>\n","      <td>MI</td>\n","      <td>19930812.0</td>\n","    </tr>\n","    <tr>\n","      <th>105</th>\n","      <td>3028</td>\n","      <td>20191231</td>\n","      <td>2019.0</td>\n","      <td>INDL</td>\n","      <td>C</td>\n","      <td>D</td>\n","      <td>STD</td>\n","      <td>CHDN</td>\n","      <td>USD</td>\n","      <td>NaN</td>\n","      <td>221.200</td>\n","      <td>2551.000</td>\n","      <td>142.500</td>\n","      <td>1491.100</td>\n","      <td>233.800</td>\n","      <td>5.500</td>\n","      <td>736.900</td>\n","      <td>0.000</td>\n","      <td>301.200</td>\n","      <td>2040.000</td>\n","      <td>137.500</td>\n","      <td>937.300</td>\n","      <td>508.300</td>\n","      <td>1329.700</td>\n","      <td>1329.700</td>\n","      <td>70.900</td>\n","      <td>NaN</td>\n","      <td>A</td>\n","      <td>Louisville</td>\n","      <td>2530</td>\n","      <td>USA</td>\n","      <td>KY</td>\n","      <td>NaN</td>\n","    </tr>\n","    <tr>\n","      <th>2239</th>\n","      <td>128978</td>\n","      <td>20191231</td>\n","      <td>2019.0</td>\n","      <td>INDL</td>\n","      <td>C</td>\n","      <td>D</td>\n","      <td>STD</td>\n","      <td>PKG</td>\n","      <td>USD</td>\n","      <td>NaN</td>\n","      <td>2478.400</td>\n","      <td>7235.800</td>\n","      <td>767.400</td>\n","      <td>2670.400</td>\n","      <td>1086.700</td>\n","      <td>15.500</td>\n","      <td>1258.900</td>\n","      <td>794.100</td>\n","      <td>723.800</td>\n","      <td>4164.800</td>\n","      <td>696.400</td>\n","      <td>3407.600</td>\n","      <td>2545.300</td>\n","      <td>6964.300</td>\n","      <td>6964.300</td>\n","      <td>93.500</td>\n","      <td>NaN</td>\n","      <td>A</td>\n","      <td>Lake Forest</td>\n","      <td>1510</td>\n","      <td>USA</td>\n","      <td>IL</td>\n","      <td>20000127.0</td>\n","    </tr>\n","    <tr>\n","      <th>1224</th>\n","      <td>25349</td>\n","      <td>20191231</td>\n","      <td>2019.0</td>\n","      <td>INDL</td>\n","      <td>C</td>\n","      <td>D</td>\n","      <td>STD</td>\n","      <td>GBT</td>\n","      <td>USD</td>\n","      <td>NaN</td>\n","      <td>627.997</td>\n","      <td>796.099</td>\n","      <td>609.969</td>\n","      <td>145.918</td>\n","      <td>-289.584</td>\n","      <td>0.352</td>\n","      <td>0.000</td>\n","      <td>1.277</td>\n","      <td>71.453</td>\n","      <td>217.405</td>\n","      <td>-266.766</td>\n","      <td>79.888</td>\n","      <td>-738.162</td>\n","      <td>2.108</td>\n","      <td>2.108</td>\n","      <td>0.894</td>\n","      <td>NaN</td>\n","      <td>A</td>\n","      <td>South San Francisco</td>\n","      <td>3520</td>\n","      <td>USA</td>\n","      <td>CA</td>\n","      <td>NaN</td>\n","    </tr>\n","    <tr>\n","      <th>1121</th>\n","      <td>23793</td>\n","      <td>20191231</td>\n","      <td>2019.0</td>\n","      <td>INDL</td>\n","      <td>C</td>\n","      <td>D</td>\n","      <td>STD</td>\n","      <td>PSB</td>\n","      <td>USD</td>\n","      <td>NaN</td>\n","      <td>NaN</td>\n","      <td>2046.443</td>\n","      <td>63.874</td>\n","      <td>0.000</td>\n","      <td>189.085</td>\n","      <td>0.155</td>\n","      <td>6.900</td>\n","      <td>0.000</td>\n","      <td>NaN</td>\n","      <td>84.632</td>\n","      <td>174.966</td>\n","      <td>NaN</td>\n","      <td>63.666</td>\n","      <td>434.338</td>\n","      <td>434.338</td>\n","      <td>0.657</td>\n","      <td>NaN</td>\n","      <td>A</td>\n","      <td>Glendale</td>\n","      <td>6010</td>\n","      <td>USA</td>\n","      <td>CA</td>\n","      <td>NaN</td>\n","    </tr>\n","    <tr>\n","      <th>2803</th>\n","      <td>196055</td>\n","      <td>20191231</td>\n","      <td>2019.0</td>\n","      <td>INDL</td>\n","      <td>C</td>\n","      <td>D</td>\n","      <td>STD</td>\n","      <td>CLVS</td>\n","      <td>USD</td>\n","      <td>NaN</td>\n","      <td>366.468</td>\n","      <td>669.604</td>\n","      <td>296.659</td>\n","      <td>709.221</td>\n","      <td>-357.595</td>\n","      <td>0.484</td>\n","      <td>125.994</td>\n","      <td>26.519</td>\n","      <td>133.084</td>\n","      <td>843.861</td>\n","      <td>-400.424</td>\n","      <td>43.428</td>\n","      <td>-2288.380</td>\n","      <td>143.006</td>\n","      <td>143.006</td>\n","      <td>19.405</td>\n","      <td>19.294</td>\n","      <td>A</td>\n","      <td>Boulder</td>\n","      <td>3520</td>\n","      <td>USA</td>\n","      <td>CO</td>\n","      <td>20111116.0</td>\n","    </tr>\n","    <tr>\n","      <th>1645</th>\n","      <td>33598</td>\n","      <td>20191231</td>\n","      <td>2019.0</td>\n","      <td>INDL</td>\n","      <td>C</td>\n","      <td>D</td>\n","      <td>STD</td>\n","      <td>USX</td>\n","      <td>USD</td>\n","      <td>NaN</td>\n","      <td>256.728</td>\n","      <td>1140.109</td>\n","      <td>5.687</td>\n","      <td>520.931</td>\n","      <td>25.239</td>\n","      <td>8.572</td>\n","      <td>84.922</td>\n","      <td>7.193</td>\n","      <td>305.491</td>\n","      <td>909.273</td>\n","      <td>-3.647</td>\n","      <td>768.401</td>\n","      <td>-20.982</td>\n","      <td>1707.361</td>\n","      <td>1707.361</td>\n","      <td>21.635</td>\n","      <td>NaN</td>\n","      <td>A</td>\n","      <td>Chattanooga</td>\n","      <td>2030</td>\n","      <td>USA</td>\n","      <td>TN</td>\n","      <td>20180614.0</td>\n","    </tr>\n","    <tr>\n","      <th>2711</th>\n","      <td>184551</td>\n","      <td>20191231</td>\n","      <td>2019.0</td>\n","      <td>INDL</td>\n","      <td>C</td>\n","      <td>D</td>\n","      <td>STD</td>\n","      <td>MXL</td>\n","      <td>USD</td>\n","      <td>NaN</td>\n","      <td>181.770</td>\n","      <td>705.791</td>\n","      <td>93.057</td>\n","      <td>216.244</td>\n","      <td>-19.421</td>\n","      <td>0.697</td>\n","      <td>426.301</td>\n","      <td>31.510</td>\n","      <td>66.562</td>\n","      <td>290.871</td>\n","      <td>-19.898</td>\n","      <td>27.591</td>\n","      <td>-114.683</td>\n","      <td>317.180</td>\n","      <td>317.180</td>\n","      <td>11.133</td>\n","      <td>NaN</td>\n","      <td>A</td>\n","      <td>Carlsbad</td>\n","      <td>4530</td>\n","      <td>USA</td>\n","      <td>CA</td>\n","      <td>20100324.0</td>\n","    </tr>\n","    <tr>\n","      <th>1436</th>\n","      <td>29508</td>\n","      <td>20190630</td>\n","      <td>2019.0</td>\n","      <td>INDL</td>\n","      <td>C</td>\n","      <td>D</td>\n","      <td>STD</td>\n","      <td>PCSB</td>\n","      <td>USD</td>\n","      <td>NaN</td>\n","      <td>NaN</td>\n","      <td>1637.579</td>\n","      <td>60.029</td>\n","      <td>46.088</td>\n","      <td>17.024</td>\n","      <td>0.182</td>\n","      <td>6.429</td>\n","      <td>1.158</td>\n","      <td>NaN</td>\n","      <td>1356.272</td>\n","      <td>8.318</td>\n","      <td>11.802</td>\n","      <td>129.410</td>\n","      <td>56.549</td>\n","      <td>56.549</td>\n","      <td>NaN</td>\n","      <td>NaN</td>\n","      <td>A</td>\n","      <td>Yorktown Heights</td>\n","      <td>4010</td>\n","      <td>USA</td>\n","      <td>NY</td>\n","      <td>20170421.0</td>\n","    </tr>\n","    <tr>\n","      <th>1960</th>\n","      <td>61635</td>\n","      <td>20191231</td>\n","      <td>2019.0</td>\n","      <td>INDL</td>\n","      <td>C</td>\n","      <td>D</td>\n","      <td>STD</td>\n","      <td>NWPX</td>\n","      <td>USD</td>\n","      <td>NaN</td>\n","      <td>195.039</td>\n","      <td>310.245</td>\n","      <td>31.014</td>\n","      <td>7.468</td>\n","      <td>30.918</td>\n","      <td>0.765</td>\n","      <td>1.231</td>\n","      <td>121.840</td>\n","      <td>41.566</td>\n","      <td>62.087</td>\n","      <td>27.902</td>\n","      <td>107.314</td>\n","      <td>127.517</td>\n","      <td>279.317</td>\n","      <td>279.317</td>\n","      <td>0.472</td>\n","      <td>NaN</td>\n","      <td>A</td>\n","      <td>Vancouver</td>\n","      <td>2010</td>\n","      <td>USA</td>\n","      <td>WA</td>\n","      <td>19951130.0</td>\n","    </tr>\n","    <tr>\n","      <th>1304</th>\n","      <td>27345</td>\n","      <td>20191231</td>\n","      <td>2019.0</td>\n","      <td>INDL</td>\n","      <td>C</td>\n","      <td>D</td>\n","      <td>STD</td>\n","      <td>BPRN</td>\n","      <td>USD</td>\n","      <td>NaN</td>\n","      <td>NaN</td>\n","      <td>1454.894</td>\n","      <td>72.598</td>\n","      <td>15.040</td>\n","      <td>19.235</td>\n","      <td>0.178</td>\n","      <td>12.616</td>\n","      <td>0.000</td>\n","      <td>NaN</td>\n","      <td>1258.969</td>\n","      <td>10.142</td>\n","      <td>25.310</td>\n","      <td>82.903</td>\n","      <td>63.109</td>\n","      <td>63.109</td>\n","      <td>NaN</td>\n","      <td>NaN</td>\n","      <td>A</td>\n","      <td>Princeton</td>\n","      <td>4010</td>\n","      <td>USA</td>\n","      <td>NJ</td>\n","      <td>20170801.0</td>\n","    </tr>\n","    <tr>\n","      <th>645</th>\n","      <td>13142</td>\n","      <td>20191231</td>\n","      <td>2019.0</td>\n","      <td>INDL</td>\n","      <td>C</td>\n","      <td>D</td>\n","      <td>STD</td>\n","      <td>CMO</td>\n","      <td>USD</td>\n","      <td>NaN</td>\n","      <td>NaN</td>\n","      <td>11520.001</td>\n","      <td>105.397</td>\n","      <td>98.392</td>\n","      <td>310.428</td>\n","      <td>0.014</td>\n","      <td>0.000</td>\n","      <td>0.000</td>\n","      <td>NaN</td>\n","      <td>10446.268</td>\n","      <td>-35.338</td>\n","      <td>NaN</td>\n","      <td>-430.640</td>\n","      <td>323.119</td>\n","      <td>323.119</td>\n","      <td>253.823</td>\n","      <td>NaN</td>\n","      <td>A</td>\n","      <td>Dallas</td>\n","      <td>4020</td>\n","      <td>USA</td>\n","      <td>TX</td>\n","      <td>19850905.0</td>\n","    </tr>\n","    <tr>\n","      <th>422</th>\n","      <td>8512</td>\n","      <td>20191231</td>\n","      <td>2019.0</td>\n","      <td>INDL</td>\n","      <td>C</td>\n","      <td>D</td>\n","      <td>STD</td>\n","      <td>PDCE</td>\n","      <td>USD</td>\n","      <td>N</td>\n","      <td>304.030</td>\n","      <td>4448.680</td>\n","      <td>0.963</td>\n","      <td>1193.005</td>\n","      <td>56.076</td>\n","      <td>0.540</td>\n","      <td>0.000</td>\n","      <td>0.000</td>\n","      <td>361.230</td>\n","      <td>2113.173</td>\n","      <td>-56.672</td>\n","      <td>4110.128</td>\n","      <td>-47.945</td>\n","      <td>1156.123</td>\n","      <td>1156.123</td>\n","      <td>84.571</td>\n","      <td>NaN</td>\n","      <td>A</td>\n","      <td>Denver</td>\n","      <td>1010</td>\n","      <td>USA</td>\n","      <td>CO</td>\n","      <td>NaN</td>\n","    </tr>\n","    <tr>\n","      <th>1087</th>\n","      <td>23083</td>\n","      <td>20190930</td>\n","      <td>2019.0</td>\n","      <td>INDL</td>\n","      <td>C</td>\n","      <td>D</td>\n","      <td>STD</td>\n","      <td>ENR</td>\n","      <td>USD</td>\n","      <td>NaN</td>\n","      <td>2036.800</td>\n","      <td>5449.600</td>\n","      <td>258.500</td>\n","      <td>3461.600</td>\n","      <td>426.900</td>\n","      <td>7.500</td>\n","      <td>2963.700</td>\n","      <td>469.300</td>\n","      <td>1069.000</td>\n","      <td>4905.800</td>\n","      <td>51.100</td>\n","      <td>362.000</td>\n","      <td>-168.800</td>\n","      <td>2494.500</td>\n","      <td>2494.500</td>\n","      <td>160.400</td>\n","      <td>NaN</td>\n","      <td>A</td>\n","      <td>Saint Louis</td>\n","      <td>3030</td>\n","      <td>USA</td>\n","      <td>MO</td>\n","      <td>NaN</td>\n","    </tr>\n","    <tr>\n","      <th>1346</th>\n","      <td>27960</td>\n","      <td>20191231</td>\n","      <td>2019.0</td>\n","      <td>INDL</td>\n","      <td>C</td>\n","      <td>D</td>\n","      <td>STD</td>\n","      <td>ALG</td>\n","      <td>USD</td>\n","      <td>NaN</td>\n","      <td>570.828</td>\n","      <td>1212.763</td>\n","      <td>42.311</td>\n","      <td>434.130</td>\n","      <td>99.839</td>\n","      <td>4.270</td>\n","      <td>404.294</td>\n","      <td>267.674</td>\n","      <td>162.874</td>\n","      <td>643.006</td>\n","      <td>62.906</td>\n","      <td>230.050</td>\n","      <td>459.482</td>\n","      <td>1119.138</td>\n","      <td>1119.138</td>\n","      <td>10.747</td>\n","      <td>NaN</td>\n","      <td>A</td>\n","      <td>Seguin</td>\n","      <td>2010</td>\n","      <td>USA</td>\n","      <td>TX</td>\n","      <td>19930318.0</td>\n","    </tr>\n","    <tr>\n","      <th>1366</th>\n","      <td>28209</td>\n","      <td>20191231</td>\n","      <td>2019.0</td>\n","      <td>INDL</td>\n","      <td>C</td>\n","      <td>D</td>\n","      <td>STD</td>\n","      <td>ANIK</td>\n","      <td>USD</td>\n","      <td>NaN</td>\n","      <td>234.306</td>\n","      <td>330.710</td>\n","      <td>184.943</td>\n","      <td>21.367</td>\n","      <td>37.496</td>\n","      <td>0.154</td>\n","      <td>15.279</td>\n","      <td>21.995</td>\n","      <td>16.277</td>\n","      <td>42.332</td>\n","      <td>27.193</td>\n","      <td>73.647</td>\n","      <td>239.528</td>\n","      <td>114.610</td>\n","      <td>114.610</td>\n","      <td>NaN</td>\n","      <td>NaN</td>\n","      <td>A</td>\n","      <td>Bedford</td>\n","      <td>3520</td>\n","      <td>USA</td>\n","      <td>MA</td>\n","      <td>19930429.0</td>\n","    </tr>\n","    <tr>\n","      <th>1926</th>\n","      <td>60900</td>\n","      <td>20191231</td>\n","      <td>2019.0</td>\n","      <td>INDL</td>\n","      <td>C</td>\n","      <td>D</td>\n","      <td>STD</td>\n","      <td>DISH</td>\n","      <td>USD</td>\n","      <td>NaN</td>\n","      <td>4015.100</td>\n","      <td>33230.935</td>\n","      <td>2860.347</td>\n","      <td>13053.024</td>\n","      <td>1878.876</td>\n","      <td>16.000</td>\n","      <td>25915.557</td>\n","      <td>322.898</td>\n","      <td>4496.091</td>\n","      <td>21114.788</td>\n","      <td>1399.512</td>\n","      <td>2850.512</td>\n","      <td>6612.284</td>\n","      <td>12807.684</td>\n","      <td>12807.684</td>\n","      <td>1003.986</td>\n","      <td>NaN</td>\n","      <td>A</td>\n","      <td>Englewood</td>\n","      <td>5020</td>\n","      <td>USA</td>\n","      <td>CO</td>\n","      <td>19950620.0</td>\n","    </tr>\n","    <tr>\n","      <th>1088</th>\n","      <td>23084</td>\n","      <td>20191231</td>\n","      <td>2019.0</td>\n","      <td>INDL</td>\n","      <td>C</td>\n","      <td>D</td>\n","      <td>STD</td>\n","      <td>BHE</td>\n","      <td>USD</td>\n","      <td>NaN</td>\n","      <td>1195.082</td>\n","      <td>1759.874</td>\n","      <td>363.956</td>\n","      <td>206.810</td>\n","      <td>49.327</td>\n","      <td>10.600</td>\n","      <td>267.909</td>\n","      <td>314.956</td>\n","      <td>459.245</td>\n","      <td>745.042</td>\n","      <td>23.425</td>\n","      <td>284.989</td>\n","      <td>499.117</td>\n","      <td>2268.095</td>\n","      <td>2268.095</td>\n","      <td>6.664</td>\n","      <td>NaN</td>\n","      <td>A</td>\n","      <td>Tempe</td>\n","      <td>4520</td>\n","      <td>USA</td>\n","      <td>AZ</td>\n","      <td>19900628.0</td>\n","    </tr>\n","    <tr>\n","      <th>176</th>\n","      <td>4383</td>\n","      <td>20190930</td>\n","      <td>2019.0</td>\n","      <td>INDL</td>\n","      <td>C</td>\n","      <td>D</td>\n","      <td>STD</td>\n","      <td>ATO</td>\n","      <td>USD</td>\n","      <td>NaN</td>\n","      <td>458.031</td>\n","      <td>13367.619</td>\n","      <td>24.550</td>\n","      <td>3529.452</td>\n","      <td>746.058</td>\n","      <td>4.776</td>\n","      <td>730.706</td>\n","      <td>136.054</td>\n","      <td>1209.440</td>\n","      <td>7617.396</td>\n","      <td>511.406</td>\n","      <td>11787.669</td>\n","      <td>2037.432</td>\n","      <td>2901.848</td>\n","      <td>2901.848</td>\n","      <td>110.796</td>\n","      <td>NaN</td>\n","      <td>A</td>\n","      <td>Dallas</td>\n","      <td>5510</td>\n","      <td>USA</td>\n","      <td>TX</td>\n","      <td>19870101.0</td>\n","    </tr>\n","    <tr>\n","      <th>1558</th>\n","      <td>32193</td>\n","      <td>20191231</td>\n","      <td>2019.0</td>\n","      <td>INDL</td>\n","      <td>C</td>\n","      <td>D</td>\n","      <td>STD</td>\n","      <td>CLXT</td>\n","      <td>USD</td>\n","      <td>NaN</td>\n","      <td>63.522</td>\n","      <td>88.098</td>\n","      <td>58.998</td>\n","      <td>18.244</td>\n","      <td>-39.673</td>\n","      <td>0.075</td>\n","      <td>0.000</td>\n","      <td>2.594</td>\n","      <td>7.196</td>\n","      <td>25.590</td>\n","      <td>-39.612</td>\n","      <td>23.212</td>\n","      <td>-122.040</td>\n","      <td>7.296</td>\n","      <td>7.296</td>\n","      <td>1.490</td>\n","      <td>NaN</td>\n","      <td>A</td>\n","      <td>Roseville</td>\n","      <td>3520</td>\n","      <td>USA</td>\n","      <td>MN</td>\n","      <td>20170720.0</td>\n","    </tr>\n","  </tbody>\n","</table>\n","</div>"],"text/plain":["       gvkey  datadate   fyear indfmt  ... ggroup  loc state     ipodate\n","1392   28742  20191231  2019.0   INDL  ...   2510  USA    MI  19930812.0\n","105     3028  20191231  2019.0   INDL  ...   2530  USA    KY         NaN\n","2239  128978  20191231  2019.0   INDL  ...   1510  USA    IL  20000127.0\n","1224   25349  20191231  2019.0   INDL  ...   3520  USA    CA         NaN\n","1121   23793  20191231  2019.0   INDL  ...   6010  USA    CA         NaN\n","2803  196055  20191231  2019.0   INDL  ...   3520  USA    CO  20111116.0\n","1645   33598  20191231  2019.0   INDL  ...   2030  USA    TN  20180614.0\n","2711  184551  20191231  2019.0   INDL  ...   4530  USA    CA  20100324.0\n","1436   29508  20190630  2019.0   INDL  ...   4010  USA    NY  20170421.0\n","1960   61635  20191231  2019.0   INDL  ...   2010  USA    WA  19951130.0\n","1304   27345  20191231  2019.0   INDL  ...   4010  USA    NJ  20170801.0\n","645    13142  20191231  2019.0   INDL  ...   4020  USA    TX  19850905.0\n","422     8512  20191231  2019.0   INDL  ...   1010  USA    CO         NaN\n","1087   23083  20190930  2019.0   INDL  ...   3030  USA    MO         NaN\n","1346   27960  20191231  2019.0   INDL  ...   2010  USA    TX  19930318.0\n","1366   28209  20191231  2019.0   INDL  ...   3520  USA    MA  19930429.0\n","1926   60900  20191231  2019.0   INDL  ...   5020  USA    CO  19950620.0\n","1088   23084  20191231  2019.0   INDL  ...   4520  USA    AZ  19900628.0\n","176     4383  20190930  2019.0   INDL  ...   5510  USA    TX  19870101.0\n","1558   32193  20191231  2019.0   INDL  ...   3520  USA    MN  20170720.0\n","\n","[20 rows x 33 columns]"]},"metadata":{"tags":[]},"execution_count":169}]},{"cell_type":"code","metadata":{"colab":{"base_uri":"https://localhost:8080/"},"id":"lSY55trc3xHN","executionInfo":{"status":"ok","timestamp":1625713107412,"user_tz":240,"elapsed":50,"user":{"displayName":"Tzu-Hua Huang","photoUrl":"https://lh3.googleusercontent.com/a-/AOh14GgbXOdZWwXGVTDuMHbeSme5C5oKAeF6I6ojY6qc=s64","userId":"01458545184256376706"}},"outputId":"1200ef7c-ef03-4171-a868-162e1e7d8f76"},"source":["# check data\n","temp2.info()"],"execution_count":170,"outputs":[{"output_type":"stream","text":["<class 'pandas.core.frame.DataFrame'>\n","RangeIndex: 2839 entries, 0 to 2838\n","Data columns (total 33 columns):\n"," #   Column    Non-Null Count  Dtype  \n","---  ------    --------------  -----  \n"," 0   gvkey     2839 non-null   int64  \n"," 1   datadate  2839 non-null   int64  \n"," 2   fyear     2838 non-null   float64\n"," 3   indfmt    2839 non-null   object \n"," 4   consol    2839 non-null   object \n"," 5   popsrc    2839 non-null   object \n"," 6   datafmt   2839 non-null   object \n"," 7   tic       2839 non-null   object \n"," 8   curcd     2838 non-null   object \n"," 9   ogm       38 non-null     object \n"," 10  act       2147 non-null   float64\n"," 11  at        2831 non-null   float64\n"," 12  che       2831 non-null   float64\n"," 13  dltt      2822 non-null   float64\n"," 14  ebit      2830 non-null   float64\n"," 15  emp       2731 non-null   float64\n"," 16  intan     2802 non-null   float64\n"," 17  invt      2793 non-null   float64\n"," 18  lct       2147 non-null   float64\n"," 19  lt        2825 non-null   float64\n"," 20  ni        2830 non-null   float64\n"," 21  ppent     2698 non-null   float64\n"," 22  re        2802 non-null   float64\n"," 23  revt      2830 non-null   float64\n"," 24  sale      2830 non-null   float64\n"," 25  xint      2367 non-null   float64\n"," 26  xintd     514 non-null    float64\n"," 27  costat    2839 non-null   object \n"," 28  city      2839 non-null   object \n"," 29  ggroup    2839 non-null   int64  \n"," 30  loc       2839 non-null   object \n"," 31  state     2748 non-null   object \n"," 32  ipodate   1708 non-null   float64\n","dtypes: float64(19), int64(3), object(11)\n","memory usage: 732.1+ KB\n"],"name":"stdout"}]},{"cell_type":"code","metadata":{"colab":{"base_uri":"https://localhost:8080/","height":258},"id":"KpU1NQqn6Utd","executionInfo":{"status":"ok","timestamp":1625713107413,"user_tz":240,"elapsed":29,"user":{"displayName":"Tzu-Hua Huang","photoUrl":"https://lh3.googleusercontent.com/a-/AOh14GgbXOdZWwXGVTDuMHbeSme5C5oKAeF6I6ojY6qc=s64","userId":"01458545184256376706"}},"outputId":"ceb0c849-c033-465f-fe50-17b734aee2f5"},"source":["# data cleaning - remove unnecessary or incomplete fields\n","temp2.drop(columns=['gvkey', 'datadate', 'fyear', 'indfmt', 'consol',\t'popsrc',\t'datafmt', \n","                    'curcd', 'ogm', 'revt', 'xintd', 'costat', 'ipodate'], inplace=True)\n","temp2.head()"],"execution_count":171,"outputs":[{"output_type":"execute_result","data":{"text/html":["<div>\n","<style scoped>\n","    .dataframe tbody tr th:only-of-type {\n","        vertical-align: middle;\n","    }\n","\n","    .dataframe tbody tr th {\n","        vertical-align: top;\n","    }\n","\n","    .dataframe thead th {\n","        text-align: right;\n","    }\n","</style>\n","<table border=\"1\" class=\"dataframe\">\n","  <thead>\n","    <tr style=\"text-align: right;\">\n","      <th></th>\n","      <th>tic</th>\n","      <th>act</th>\n","      <th>at</th>\n","      <th>che</th>\n","      <th>dltt</th>\n","      <th>ebit</th>\n","      <th>emp</th>\n","      <th>intan</th>\n","      <th>invt</th>\n","      <th>lct</th>\n","      <th>lt</th>\n","      <th>ni</th>\n","      <th>ppent</th>\n","      <th>re</th>\n","      <th>sale</th>\n","      <th>xint</th>\n","      <th>city</th>\n","      <th>ggroup</th>\n","      <th>loc</th>\n","      <th>state</th>\n","    </tr>\n","  </thead>\n","  <tbody>\n","    <tr>\n","      <th>0</th>\n","      <td>AIR</td>\n","      <td>1438.700</td>\n","      <td>2079.000</td>\n","      <td>424.700</td>\n","      <td>670.900</td>\n","      <td>108.400</td>\n","      <td>5.40</td>\n","      <td>125.800</td>\n","      <td>692.700</td>\n","      <td>383.100</td>\n","      <td>1176.400</td>\n","      <td>4.400</td>\n","      <td>437.100</td>\n","      <td>661.400</td>\n","      <td>2089.300</td>\n","      <td>9.300</td>\n","      <td>Wood Dale</td>\n","      <td>2010</td>\n","      <td>USA</td>\n","      <td>IL</td>\n","    </tr>\n","    <tr>\n","      <th>1</th>\n","      <td>AAL</td>\n","      <td>8206.000</td>\n","      <td>59995.000</td>\n","      <td>3984.000</td>\n","      <td>28875.000</td>\n","      <td>3706.000</td>\n","      <td>133.70</td>\n","      <td>6175.000</td>\n","      <td>1851.000</td>\n","      <td>18311.000</td>\n","      <td>60113.000</td>\n","      <td>1686.000</td>\n","      <td>43732.000</td>\n","      <td>-4067.000</td>\n","      <td>45768.000</td>\n","      <td>1095.000</td>\n","      <td>Fort Worth</td>\n","      <td>2030</td>\n","      <td>USA</td>\n","      <td>TX</td>\n","    </tr>\n","    <tr>\n","      <th>2</th>\n","      <td>CECE</td>\n","      <td>179.498</td>\n","      <td>408.637</td>\n","      <td>36.958</td>\n","      <td>81.457</td>\n","      <td>19.622</td>\n","      <td>0.83</td>\n","      <td>197.594</td>\n","      <td>55.383</td>\n","      <td>115.188</td>\n","      <td>215.620</td>\n","      <td>17.707</td>\n","      <td>28.881</td>\n","      <td>-60.849</td>\n","      <td>341.869</td>\n","      <td>4.997</td>\n","      <td>Dallas</td>\n","      <td>2020</td>\n","      <td>USA</td>\n","      <td>TX</td>\n","    </tr>\n","    <tr>\n","      <th>3</th>\n","      <td>PNW</td>\n","      <td>1030.030</td>\n","      <td>18479.247</td>\n","      <td>10.283</td>\n","      <td>4884.430</td>\n","      <td>671.960</td>\n","      <td>6.21</td>\n","      <td>290.564</td>\n","      <td>345.920</td>\n","      <td>2078.365</td>\n","      <td>12926.059</td>\n","      <td>538.320</td>\n","      <td>14377.787</td>\n","      <td>2780.514</td>\n","      <td>3471.209</td>\n","      <td>235.251</td>\n","      <td>Phoenix</td>\n","      <td>5510</td>\n","      <td>USA</td>\n","      <td>AZ</td>\n","    </tr>\n","    <tr>\n","      <th>4</th>\n","      <td>PRG</td>\n","      <td>NaN</td>\n","      <td>3297.800</td>\n","      <td>57.755</td>\n","      <td>626.530</td>\n","      <td>313.931</td>\n","      <td>12.10</td>\n","      <td>927.378</td>\n","      <td>14.000</td>\n","      <td>NaN</td>\n","      <td>1560.541</td>\n","      <td>31.472</td>\n","      <td>1986.294</td>\n","      <td>2029.594</td>\n","      <td>3947.656</td>\n","      <td>16.967</td>\n","      <td>Draper</td>\n","      <td>4020</td>\n","      <td>USA</td>\n","      <td>UT</td>\n","    </tr>\n","  </tbody>\n","</table>\n","</div>"],"text/plain":["    tic       act         at       che  ...        city  ggroup  loc  state\n","0   AIR  1438.700   2079.000   424.700  ...   Wood Dale    2010  USA     IL\n","1   AAL  8206.000  59995.000  3984.000  ...  Fort Worth    2030  USA     TX\n","2  CECE   179.498    408.637    36.958  ...      Dallas    2020  USA     TX\n","3   PNW  1030.030  18479.247    10.283  ...     Phoenix    5510  USA     AZ\n","4   PRG       NaN   3297.800    57.755  ...      Draper    4020  USA     UT\n","\n","[5 rows x 20 columns]"]},"metadata":{"tags":[]},"execution_count":171}]},{"cell_type":"code","metadata":{"colab":{"base_uri":"https://localhost:8080/","height":456},"id":"AqG_6De5xUY_","executionInfo":{"status":"ok","timestamp":1625713107419,"user_tz":240,"elapsed":32,"user":{"displayName":"Tzu-Hua Huang","photoUrl":"https://lh3.googleusercontent.com/a-/AOh14GgbXOdZWwXGVTDuMHbeSme5C5oKAeF6I6ojY6qc=s64","userId":"01458545184256376706"}},"outputId":"0e582e23-dc0d-4506-ad83-b3d94d9fc542"},"source":["# merge data set temp1 anf temp2\n","df = temp1.merge(temp2, left_on='TICKER', right_on='tic')\n","df"],"execution_count":172,"outputs":[{"output_type":"execute_result","data":{"text/html":["<div>\n","<style scoped>\n","    .dataframe tbody tr th:only-of-type {\n","        vertical-align: middle;\n","    }\n","\n","    .dataframe tbody tr th {\n","        vertical-align: top;\n","    }\n","\n","    .dataframe thead th {\n","        text-align: right;\n","    }\n","</style>\n","<table border=\"1\" class=\"dataframe\">\n","  <thead>\n","    <tr style=\"text-align: right;\">\n","      <th></th>\n","      <th>TICKER</th>\n","      <th>RetEarly2020</th>\n","      <th>RetLate2020</th>\n","      <th>tic</th>\n","      <th>act</th>\n","      <th>at</th>\n","      <th>che</th>\n","      <th>dltt</th>\n","      <th>ebit</th>\n","      <th>emp</th>\n","      <th>intan</th>\n","      <th>invt</th>\n","      <th>lct</th>\n","      <th>lt</th>\n","      <th>ni</th>\n","      <th>ppent</th>\n","      <th>re</th>\n","      <th>sale</th>\n","      <th>xint</th>\n","      <th>city</th>\n","      <th>ggroup</th>\n","      <th>loc</th>\n","      <th>state</th>\n","    </tr>\n","  </thead>\n","  <tbody>\n","    <tr>\n","      <th>0</th>\n","      <td>JJSF</td>\n","      <td>-0.340234</td>\n","      <td>0.300340</td>\n","      <td>JJSF</td>\n","      <td>506.357</td>\n","      <td>1019.339</td>\n","      <td>243.486</td>\n","      <td>0.718</td>\n","      <td>116.956</td>\n","      <td>4.600</td>\n","      <td>157.433</td>\n","      <td>116.165</td>\n","      <td>121.234</td>\n","      <td>185.588</td>\n","      <td>94.819</td>\n","      <td>253.448</td>\n","      <td>788.007</td>\n","      <td>1186.487</td>\n","      <td>NaN</td>\n","      <td>Pennsauken</td>\n","      <td>3020</td>\n","      <td>USA</td>\n","      <td>NJ</td>\n","    </tr>\n","    <tr>\n","      <th>1</th>\n","      <td>PLXS</td>\n","      <td>-0.290876</td>\n","      <td>0.433469</td>\n","      <td>PLXS</td>\n","      <td>1538.291</td>\n","      <td>2000.883</td>\n","      <td>226.254</td>\n","      <td>187.278</td>\n","      <td>143.733</td>\n","      <td>19.000</td>\n","      <td>NaN</td>\n","      <td>700.938</td>\n","      <td>865.503</td>\n","      <td>1135.307</td>\n","      <td>108.616</td>\n","      <td>384.224</td>\n","      <td>1160.893</td>\n","      <td>3164.434</td>\n","      <td>12.853</td>\n","      <td>Neenah</td>\n","      <td>4520</td>\n","      <td>USA</td>\n","      <td>WI</td>\n","    </tr>\n","    <tr>\n","      <th>2</th>\n","      <td>HNGR</td>\n","      <td>-0.435712</td>\n","      <td>0.411425</td>\n","      <td>HNGR</td>\n","      <td>315.655</td>\n","      <td>842.253</td>\n","      <td>74.419</td>\n","      <td>578.539</td>\n","      <td>66.928</td>\n","      <td>4.800</td>\n","      <td>250.196</td>\n","      <td>68.204</td>\n","      <td>208.406</td>\n","      <td>832.749</td>\n","      <td>27.525</td>\n","      <td>196.104</td>\n","      <td>-344.502</td>\n","      <td>1098.046</td>\n","      <td>34.258</td>\n","      <td>Austin</td>\n","      <td>3510</td>\n","      <td>USA</td>\n","      <td>TX</td>\n","    </tr>\n","    <tr>\n","      <th>3</th>\n","      <td>ORCL</td>\n","      <td>-0.083596</td>\n","      <td>0.356166</td>\n","      <td>ORCL</td>\n","      <td>52140.000</td>\n","      <td>115438.000</td>\n","      <td>43057.000</td>\n","      <td>70726.000</td>\n","      <td>14139.000</td>\n","      <td>135.000</td>\n","      <td>47507.000</td>\n","      <td>211.000</td>\n","      <td>17200.000</td>\n","      <td>102721.000</td>\n","      <td>10135.000</td>\n","      <td>8244.000</td>\n","      <td>-14412.000</td>\n","      <td>39068.000</td>\n","      <td>1995.000</td>\n","      <td>Austin</td>\n","      <td>4510</td>\n","      <td>USA</td>\n","      <td>TX</td>\n","    </tr>\n","    <tr>\n","      <th>4</th>\n","      <td>MSFT</td>\n","      <td>0.003212</td>\n","      <td>0.421139</td>\n","      <td>MSFT</td>\n","      <td>175552.000</td>\n","      <td>286556.000</td>\n","      <td>133819.000</td>\n","      <td>79107.000</td>\n","      <td>42959.000</td>\n","      <td>144.000</td>\n","      <td>49776.000</td>\n","      <td>2063.000</td>\n","      <td>69420.000</td>\n","      <td>184226.000</td>\n","      <td>39240.000</td>\n","      <td>43856.000</td>\n","      <td>23810.000</td>\n","      <td>125843.000</td>\n","      <td>2686.000</td>\n","      <td>Redmond</td>\n","      <td>4510</td>\n","      <td>USA</td>\n","      <td>WA</td>\n","    </tr>\n","    <tr>\n","      <th>...</th>\n","      <td>...</td>\n","      <td>...</td>\n","      <td>...</td>\n","      <td>...</td>\n","      <td>...</td>\n","      <td>...</td>\n","      <td>...</td>\n","      <td>...</td>\n","      <td>...</td>\n","      <td>...</td>\n","      <td>...</td>\n","      <td>...</td>\n","      <td>...</td>\n","      <td>...</td>\n","      <td>...</td>\n","      <td>...</td>\n","      <td>...</td>\n","      <td>...</td>\n","      <td>...</td>\n","      <td>...</td>\n","      <td>...</td>\n","      <td>...</td>\n","      <td>...</td>\n","    </tr>\n","    <tr>\n","      <th>2661</th>\n","      <td>SIX</td>\n","      <td>-0.716471</td>\n","      <td>1.719300</td>\n","      <td>SIX</td>\n","      <td>376.240</td>\n","      <td>2882.540</td>\n","      <td>174.179</td>\n","      <td>2455.033</td>\n","      <td>424.564</td>\n","      <td>50.450</td>\n","      <td>1004.830</td>\n","      <td>55.851</td>\n","      <td>339.732</td>\n","      <td>3069.400</td>\n","      <td>179.065</td>\n","      <td>1485.124</td>\n","      <td>-1784.457</td>\n","      <td>1480.083</td>\n","      <td>114.703</td>\n","      <td>Arlington</td>\n","      <td>2530</td>\n","      <td>USA</td>\n","      <td>TX</td>\n","    </tr>\n","    <tr>\n","      <th>2662</th>\n","      <td>VPG</td>\n","      <td>-0.409412</td>\n","      <td>0.567728</td>\n","      <td>VPG</td>\n","      <td>212.561</td>\n","      <td>370.413</td>\n","      <td>86.910</td>\n","      <td>5.828</td>\n","      <td>33.249</td>\n","      <td>2.400</td>\n","      <td>69.216</td>\n","      <td>66.895</td>\n","      <td>88.911</td>\n","      <td>128.661</td>\n","      <td>22.188</td>\n","      <td>69.961</td>\n","      <td>51.585</td>\n","      <td>283.958</td>\n","      <td>1.507</td>\n","      <td>Malvern</td>\n","      <td>4520</td>\n","      <td>USA</td>\n","      <td>PA</td>\n","    </tr>\n","    <tr>\n","      <th>2663</th>\n","      <td>FN</td>\n","      <td>-0.158545</td>\n","      <td>0.422103</td>\n","      <td>FN</td>\n","      <td>1023.835</td>\n","      <td>1255.318</td>\n","      <td>437.332</td>\n","      <td>57.790</td>\n","      <td>124.999</td>\n","      <td>11.670</td>\n","      <td>7.592</td>\n","      <td>293.612</td>\n","      <td>313.048</td>\n","      <td>392.219</td>\n","      <td>120.955</td>\n","      <td>210.686</td>\n","      <td>752.197</td>\n","      <td>1584.335</td>\n","      <td>5.381</td>\n","      <td>George Town</td>\n","      <td>4520</td>\n","      <td>CYM</td>\n","      <td>NaN</td>\n","    </tr>\n","    <tr>\n","      <th>2664</th>\n","      <td>CBOE</td>\n","      <td>-0.253901</td>\n","      <td>0.056516</td>\n","      <td>CBOE</td>\n","      <td>607.600</td>\n","      <td>5113.900</td>\n","      <td>300.300</td>\n","      <td>914.300</td>\n","      <td>585.700</td>\n","      <td>0.823</td>\n","      <td>4301.700</td>\n","      <td>0.000</td>\n","      <td>281.600</td>\n","      <td>1758.300</td>\n","      <td>374.400</td>\n","      <td>100.400</td>\n","      <td>1550.200</td>\n","      <td>2496.100</td>\n","      <td>37.800</td>\n","      <td>Chicago</td>\n","      <td>4020</td>\n","      <td>USA</td>\n","      <td>IL</td>\n","    </tr>\n","    <tr>\n","      <th>2665</th>\n","      <td>TSLA</td>\n","      <td>0.252599</td>\n","      <td>5.733495</td>\n","      <td>TSLA</td>\n","      <td>12103.000</td>\n","      <td>34309.000</td>\n","      <td>6514.000</td>\n","      <td>12627.000</td>\n","      <td>80.000</td>\n","      <td>48.016</td>\n","      <td>537.000</td>\n","      <td>3552.000</td>\n","      <td>10667.000</td>\n","      <td>26199.000</td>\n","      <td>-862.000</td>\n","      <td>20199.000</td>\n","      <td>-6119.000</td>\n","      <td>24578.000</td>\n","      <td>716.000</td>\n","      <td>Palo Alto</td>\n","      <td>2510</td>\n","      <td>USA</td>\n","      <td>CA</td>\n","    </tr>\n","  </tbody>\n","</table>\n","<p>2666 rows × 23 columns</p>\n","</div>"],"text/plain":["     TICKER  RetEarly2020  RetLate2020   tic  ...         city  ggroup  loc  state\n","0      JJSF     -0.340234     0.300340  JJSF  ...   Pennsauken    3020  USA     NJ\n","1      PLXS     -0.290876     0.433469  PLXS  ...       Neenah    4520  USA     WI\n","2      HNGR     -0.435712     0.411425  HNGR  ...       Austin    3510  USA     TX\n","3      ORCL     -0.083596     0.356166  ORCL  ...       Austin    4510  USA     TX\n","4      MSFT      0.003212     0.421139  MSFT  ...      Redmond    4510  USA     WA\n","...     ...           ...          ...   ...  ...          ...     ...  ...    ...\n","2661    SIX     -0.716471     1.719300   SIX  ...    Arlington    2530  USA     TX\n","2662    VPG     -0.409412     0.567728   VPG  ...      Malvern    4520  USA     PA\n","2663     FN     -0.158545     0.422103    FN  ...  George Town    4520  CYM    NaN\n","2664   CBOE     -0.253901     0.056516  CBOE  ...      Chicago    4020  USA     IL\n","2665   TSLA      0.252599     5.733495  TSLA  ...    Palo Alto    2510  USA     CA\n","\n","[2666 rows x 23 columns]"]},"metadata":{"tags":[]},"execution_count":172}]},{"cell_type":"code","metadata":{"colab":{"base_uri":"https://localhost:8080/","height":258},"id":"BTYjXh709G5Z","executionInfo":{"status":"ok","timestamp":1625713107707,"user_tz":240,"elapsed":317,"user":{"displayName":"Tzu-Hua Huang","photoUrl":"https://lh3.googleusercontent.com/a-/AOh14GgbXOdZWwXGVTDuMHbeSme5C5oKAeF6I6ojY6qc=s64","userId":"01458545184256376706"}},"outputId":"32814f1c-ab64-47bf-ebea-c8b6678d6382"},"source":["# clean and rename columns\n","df.drop(columns=['tic'], inplace=True)\n","df.rename(columns={'act':'Current Assets', 'at':'Total Assets', 'che':'Cash', 'ebit':'EBIT', \n","                   'dltt':'Long-term Debt', 'emp':'#Employees', 'intan':'Intangible Assets', \n","                   'invt':'Inventories', 'lct':'Current Liabil', 'lt':'Total Liabil', 'ni':'NI', \n","                   'ppent':'PP&E', 're':'Retained Earnings', 'sale':'Sales', 'xint':'Interest Expenses',\n","                   'city':'City', 'ggroup':'GGROUP', 'loc':'Location', 'state':'State'}, inplace=True)\n","df.head()"],"execution_count":173,"outputs":[{"output_type":"execute_result","data":{"text/html":["<div>\n","<style scoped>\n","    .dataframe tbody tr th:only-of-type {\n","        vertical-align: middle;\n","    }\n","\n","    .dataframe tbody tr th {\n","        vertical-align: top;\n","    }\n","\n","    .dataframe thead th {\n","        text-align: right;\n","    }\n","</style>\n","<table border=\"1\" class=\"dataframe\">\n","  <thead>\n","    <tr style=\"text-align: right;\">\n","      <th></th>\n","      <th>TICKER</th>\n","      <th>RetEarly2020</th>\n","      <th>RetLate2020</th>\n","      <th>Current Assets</th>\n","      <th>Total Assets</th>\n","      <th>Cash</th>\n","      <th>Long-term Debt</th>\n","      <th>EBIT</th>\n","      <th>#Employees</th>\n","      <th>Intangible Assets</th>\n","      <th>Inventories</th>\n","      <th>Current Liabil</th>\n","      <th>Total Liabil</th>\n","      <th>NI</th>\n","      <th>PP&amp;E</th>\n","      <th>Retained Earnings</th>\n","      <th>Sales</th>\n","      <th>Interest Expenses</th>\n","      <th>City</th>\n","      <th>GGROUP</th>\n","      <th>Location</th>\n","      <th>State</th>\n","    </tr>\n","  </thead>\n","  <tbody>\n","    <tr>\n","      <th>0</th>\n","      <td>JJSF</td>\n","      <td>-0.340234</td>\n","      <td>0.300340</td>\n","      <td>506.357</td>\n","      <td>1019.339</td>\n","      <td>243.486</td>\n","      <td>0.718</td>\n","      <td>116.956</td>\n","      <td>4.6</td>\n","      <td>157.433</td>\n","      <td>116.165</td>\n","      <td>121.234</td>\n","      <td>185.588</td>\n","      <td>94.819</td>\n","      <td>253.448</td>\n","      <td>788.007</td>\n","      <td>1186.487</td>\n","      <td>NaN</td>\n","      <td>Pennsauken</td>\n","      <td>3020</td>\n","      <td>USA</td>\n","      <td>NJ</td>\n","    </tr>\n","    <tr>\n","      <th>1</th>\n","      <td>PLXS</td>\n","      <td>-0.290876</td>\n","      <td>0.433469</td>\n","      <td>1538.291</td>\n","      <td>2000.883</td>\n","      <td>226.254</td>\n","      <td>187.278</td>\n","      <td>143.733</td>\n","      <td>19.0</td>\n","      <td>NaN</td>\n","      <td>700.938</td>\n","      <td>865.503</td>\n","      <td>1135.307</td>\n","      <td>108.616</td>\n","      <td>384.224</td>\n","      <td>1160.893</td>\n","      <td>3164.434</td>\n","      <td>12.853</td>\n","      <td>Neenah</td>\n","      <td>4520</td>\n","      <td>USA</td>\n","      <td>WI</td>\n","    </tr>\n","    <tr>\n","      <th>2</th>\n","      <td>HNGR</td>\n","      <td>-0.435712</td>\n","      <td>0.411425</td>\n","      <td>315.655</td>\n","      <td>842.253</td>\n","      <td>74.419</td>\n","      <td>578.539</td>\n","      <td>66.928</td>\n","      <td>4.8</td>\n","      <td>250.196</td>\n","      <td>68.204</td>\n","      <td>208.406</td>\n","      <td>832.749</td>\n","      <td>27.525</td>\n","      <td>196.104</td>\n","      <td>-344.502</td>\n","      <td>1098.046</td>\n","      <td>34.258</td>\n","      <td>Austin</td>\n","      <td>3510</td>\n","      <td>USA</td>\n","      <td>TX</td>\n","    </tr>\n","    <tr>\n","      <th>3</th>\n","      <td>ORCL</td>\n","      <td>-0.083596</td>\n","      <td>0.356166</td>\n","      <td>52140.000</td>\n","      <td>115438.000</td>\n","      <td>43057.000</td>\n","      <td>70726.000</td>\n","      <td>14139.000</td>\n","      <td>135.0</td>\n","      <td>47507.000</td>\n","      <td>211.000</td>\n","      <td>17200.000</td>\n","      <td>102721.000</td>\n","      <td>10135.000</td>\n","      <td>8244.000</td>\n","      <td>-14412.000</td>\n","      <td>39068.000</td>\n","      <td>1995.000</td>\n","      <td>Austin</td>\n","      <td>4510</td>\n","      <td>USA</td>\n","      <td>TX</td>\n","    </tr>\n","    <tr>\n","      <th>4</th>\n","      <td>MSFT</td>\n","      <td>0.003212</td>\n","      <td>0.421139</td>\n","      <td>175552.000</td>\n","      <td>286556.000</td>\n","      <td>133819.000</td>\n","      <td>79107.000</td>\n","      <td>42959.000</td>\n","      <td>144.0</td>\n","      <td>49776.000</td>\n","      <td>2063.000</td>\n","      <td>69420.000</td>\n","      <td>184226.000</td>\n","      <td>39240.000</td>\n","      <td>43856.000</td>\n","      <td>23810.000</td>\n","      <td>125843.000</td>\n","      <td>2686.000</td>\n","      <td>Redmond</td>\n","      <td>4510</td>\n","      <td>USA</td>\n","      <td>WA</td>\n","    </tr>\n","  </tbody>\n","</table>\n","</div>"],"text/plain":["  TICKER  RetEarly2020  RetLate2020  ...  GGROUP  Location  State\n","0   JJSF     -0.340234     0.300340  ...    3020       USA     NJ\n","1   PLXS     -0.290876     0.433469  ...    4520       USA     WI\n","2   HNGR     -0.435712     0.411425  ...    3510       USA     TX\n","3   ORCL     -0.083596     0.356166  ...    4510       USA     TX\n","4   MSFT      0.003212     0.421139  ...    4510       USA     WA\n","\n","[5 rows x 22 columns]"]},"metadata":{"tags":[]},"execution_count":173}]},{"cell_type":"markdown","metadata":{"id":"7AcL-maxu0_U"},"source":["- According [WRDS Compustat North America User’s Guide](http://sites.bu.edu/qm222projectcourse/files/2014/08/compustat_users_guide-2003.pdf), the unit of `#Employees` is 'Thousands'. Adjust the column values by multiplying the column values by 1000.\n","- Add new column `Total Equity` for later use."]},{"cell_type":"code","metadata":{"colab":{"base_uri":"https://localhost:8080/"},"id":"AlN4KLIDvxiM","executionInfo":{"status":"ok","timestamp":1625713107708,"user_tz":240,"elapsed":75,"user":{"displayName":"Tzu-Hua Huang","photoUrl":"https://lh3.googleusercontent.com/a-/AOh14GgbXOdZWwXGVTDuMHbeSme5C5oKAeF6I6ojY6qc=s64","userId":"01458545184256376706"}},"outputId":"1de50389-a333-4491-96a7-bcca1842336d"},"source":["# fix column values to desirable unit\n","df['#Employees'] = df['#Employees'] * 1000\n","\n","# add new column - Total Equity\n","df['Total Equity'] = df['Total Assets'] - df['Total Liabil']\n","df.info()"],"execution_count":174,"outputs":[{"output_type":"stream","text":["<class 'pandas.core.frame.DataFrame'>\n","Int64Index: 2666 entries, 0 to 2665\n","Data columns (total 23 columns):\n"," #   Column             Non-Null Count  Dtype  \n","---  ------             --------------  -----  \n"," 0   TICKER             2666 non-null   object \n"," 1   RetEarly2020       2666 non-null   float64\n"," 2   RetLate2020        2666 non-null   float64\n"," 3   Current Assets     2003 non-null   float64\n"," 4   Total Assets       2664 non-null   float64\n"," 5   Cash               2664 non-null   float64\n"," 6   Long-term Debt     2657 non-null   float64\n"," 7   EBIT               2664 non-null   float64\n"," 8   #Employees         2647 non-null   float64\n"," 9   Intangible Assets  2635 non-null   float64\n"," 10  Inventories        2626 non-null   float64\n"," 11  Current Liabil     2003 non-null   float64\n"," 12  Total Liabil       2658 non-null   float64\n"," 13  NI                 2664 non-null   float64\n"," 14  PP&E               2534 non-null   float64\n"," 15  Retained Earnings  2656 non-null   float64\n"," 16  Sales              2664 non-null   float64\n"," 17  Interest Expenses  2222 non-null   float64\n"," 18  City               2666 non-null   object \n"," 19  GGROUP             2666 non-null   int64  \n"," 20  Location           2666 non-null   object \n"," 21  State              2582 non-null   object \n"," 22  Total Equity       2658 non-null   float64\n","dtypes: float64(18), int64(1), object(4)\n","memory usage: 499.9+ KB\n"],"name":"stdout"}]},{"cell_type":"code","metadata":{"colab":{"base_uri":"https://localhost:8080/","height":760},"id":"sxDDpfsB6Od6","executionInfo":{"status":"ok","timestamp":1625713107709,"user_tz":240,"elapsed":65,"user":{"displayName":"Tzu-Hua Huang","photoUrl":"https://lh3.googleusercontent.com/a-/AOh14GgbXOdZWwXGVTDuMHbeSme5C5oKAeF6I6ojY6qc=s64","userId":"01458545184256376706"}},"outputId":"7abad013-a0fe-4617-a79e-e79e16f2e331"},"source":["df.describe().T"],"execution_count":175,"outputs":[{"output_type":"execute_result","data":{"text/html":["<div>\n","<style scoped>\n","    .dataframe tbody tr th:only-of-type {\n","        vertical-align: middle;\n","    }\n","\n","    .dataframe tbody tr th {\n","        vertical-align: top;\n","    }\n","\n","    .dataframe thead th {\n","        text-align: right;\n","    }\n","</style>\n","<table border=\"1\" class=\"dataframe\">\n","  <thead>\n","    <tr style=\"text-align: right;\">\n","      <th></th>\n","      <th>count</th>\n","      <th>mean</th>\n","      <th>std</th>\n","      <th>min</th>\n","      <th>25%</th>\n","      <th>50%</th>\n","      <th>75%</th>\n","      <th>max</th>\n","    </tr>\n","  </thead>\n","  <tbody>\n","    <tr>\n","      <th>RetEarly2020</th>\n","      <td>2666.0</td>\n","      <td>-0.290915</td>\n","      <td>0.313196</td>\n","      <td>-0.908665</td>\n","      <td>-0.440401</td>\n","      <td>-0.312152</td>\n","      <td>-0.171201</td>\n","      <td>7.512067e+00</td>\n","    </tr>\n","    <tr>\n","      <th>RetLate2020</th>\n","      <td>2666.0</td>\n","      <td>0.804400</td>\n","      <td>1.074286</td>\n","      <td>-0.858458</td>\n","      <td>0.252418</td>\n","      <td>0.547064</td>\n","      <td>1.000302</td>\n","      <td>1.899385e+01</td>\n","    </tr>\n","    <tr>\n","      <th>Current Assets</th>\n","      <td>2003.0</td>\n","      <td>2716.237148</td>\n","      <td>9611.607647</td>\n","      <td>1.584000</td>\n","      <td>211.649500</td>\n","      <td>553.918000</td>\n","      <td>1658.592000</td>\n","      <td>1.755520e+05</td>\n","    </tr>\n","    <tr>\n","      <th>Total Assets</th>\n","      <td>2664.0</td>\n","      <td>17351.452124</td>\n","      <td>101497.142116</td>\n","      <td>2.215000</td>\n","      <td>729.553750</td>\n","      <td>2354.388000</td>\n","      <td>7813.852500</td>\n","      <td>2.687379e+06</td>\n","    </tr>\n","    <tr>\n","      <th>Cash</th>\n","      <td>2664.0</td>\n","      <td>1943.994080</td>\n","      <td>18392.290725</td>\n","      <td>0.000000</td>\n","      <td>63.961250</td>\n","      <td>176.490500</td>\n","      <td>523.115000</td>\n","      <td>5.127880e+05</td>\n","    </tr>\n","    <tr>\n","      <th>Long-term Debt</th>\n","      <td>2657.0</td>\n","      <td>3439.395660</td>\n","      <td>13418.133870</td>\n","      <td>0.000000</td>\n","      <td>60.818000</td>\n","      <td>468.300000</td>\n","      <td>2154.100000</td>\n","      <td>2.617200e+05</td>\n","    </tr>\n","    <tr>\n","      <th>EBIT</th>\n","      <td>2664.0</td>\n","      <td>802.302467</td>\n","      <td>3251.873392</td>\n","      <td>-8596.000000</td>\n","      <td>17.606250</td>\n","      <td>113.542000</td>\n","      <td>459.727000</td>\n","      <td>6.393000e+04</td>\n","    </tr>\n","    <tr>\n","      <th>#Employees</th>\n","      <td>2647.0</td>\n","      <td>15090.848130</td>\n","      <td>60891.497973</td>\n","      <td>0.000000</td>\n","      <td>448.500000</td>\n","      <td>2290.000000</td>\n","      <td>9500.000000</td>\n","      <td>2.200000e+06</td>\n","    </tr>\n","    <tr>\n","      <th>Intangible Assets</th>\n","      <td>2635.0</td>\n","      <td>2477.931363</td>\n","      <td>10750.593625</td>\n","      <td>0.000000</td>\n","      <td>14.664500</td>\n","      <td>187.907000</td>\n","      <td>1100.097500</td>\n","      <td>3.038580e+05</td>\n","    </tr>\n","    <tr>\n","      <th>Inventories</th>\n","      <td>2626.0</td>\n","      <td>948.039702</td>\n","      <td>11731.256877</td>\n","      <td>0.000000</td>\n","      <td>0.000000</td>\n","      <td>21.925500</td>\n","      <td>251.135000</td>\n","      <td>4.023900e+05</td>\n","    </tr>\n","    <tr>\n","      <th>Current Liabil</th>\n","      <td>2003.0</td>\n","      <td>2107.014083</td>\n","      <td>7516.854499</td>\n","      <td>0.328000</td>\n","      <td>78.357000</td>\n","      <td>290.000000</td>\n","      <td>1074.699000</td>\n","      <td>1.057180e+05</td>\n","    </tr>\n","    <tr>\n","      <th>Total Liabil</th>\n","      <td>2658.0</td>\n","      <td>13558.881777</td>\n","      <td>90165.735362</td>\n","      <td>0.478000</td>\n","      <td>330.327000</td>\n","      <td>1517.426500</td>\n","      <td>5332.281000</td>\n","      <td>2.426049e+06</td>\n","    </tr>\n","    <tr>\n","      <th>NI</th>\n","      <td>2664.0</td>\n","      <td>489.509873</td>\n","      <td>2362.924277</td>\n","      <td>-10137.000000</td>\n","      <td>-8.673250</td>\n","      <td>53.703000</td>\n","      <td>263.813000</td>\n","      <td>5.525600e+04</td>\n","    </tr>\n","    <tr>\n","      <th>PP&amp;E</th>\n","      <td>2534.0</td>\n","      <td>2655.320498</td>\n","      <td>10846.385400</td>\n","      <td>0.000000</td>\n","      <td>43.419750</td>\n","      <td>231.139000</td>\n","      <td>1090.268000</td>\n","      <td>2.596510e+05</td>\n","    </tr>\n","    <tr>\n","      <th>Retained Earnings</th>\n","      <td>2656.0</td>\n","      <td>2312.222038</td>\n","      <td>13695.937385</td>\n","      <td>-69382.300000</td>\n","      <td>-198.992500</td>\n","      <td>121.628500</td>\n","      <td>917.807500</td>\n","      <td>4.018480e+05</td>\n","    </tr>\n","    <tr>\n","      <th>Sales</th>\n","      <td>2664.0</td>\n","      <td>5907.023291</td>\n","      <td>21687.210521</td>\n","      <td>0.000000</td>\n","      <td>262.486750</td>\n","      <td>985.912000</td>\n","      <td>3477.726000</td>\n","      <td>5.214260e+05</td>\n","    </tr>\n","    <tr>\n","      <th>Interest Expenses</th>\n","      <td>2222.0</td>\n","      <td>171.782209</td>\n","      <td>664.963498</td>\n","      <td>0.000000</td>\n","      <td>5.017750</td>\n","      <td>31.959000</td>\n","      <td>129.880000</td>\n","      <td>1.653000e+04</td>\n","    </tr>\n","    <tr>\n","      <th>GGROUP</th>\n","      <td>2666.0</td>\n","      <td>3482.070518</td>\n","      <td>1275.214785</td>\n","      <td>1010.000000</td>\n","      <td>2520.000000</td>\n","      <td>3520.000000</td>\n","      <td>4510.000000</td>\n","      <td>6.010000e+03</td>\n","    </tr>\n","    <tr>\n","      <th>Total Equity</th>\n","      <td>2658.0</td>\n","      <td>3830.273234</td>\n","      <td>14341.572405</td>\n","      <td>-9599.000000</td>\n","      <td>241.808000</td>\n","      <td>743.137000</td>\n","      <td>2339.679750</td>\n","      <td>2.648100e+05</td>\n","    </tr>\n","  </tbody>\n","</table>\n","</div>"],"text/plain":["                    count          mean  ...          75%           max\n","RetEarly2020       2666.0     -0.290915  ...    -0.171201  7.512067e+00\n","RetLate2020        2666.0      0.804400  ...     1.000302  1.899385e+01\n","Current Assets     2003.0   2716.237148  ...  1658.592000  1.755520e+05\n","Total Assets       2664.0  17351.452124  ...  7813.852500  2.687379e+06\n","Cash               2664.0   1943.994080  ...   523.115000  5.127880e+05\n","Long-term Debt     2657.0   3439.395660  ...  2154.100000  2.617200e+05\n","EBIT               2664.0    802.302467  ...   459.727000  6.393000e+04\n","#Employees         2647.0  15090.848130  ...  9500.000000  2.200000e+06\n","Intangible Assets  2635.0   2477.931363  ...  1100.097500  3.038580e+05\n","Inventories        2626.0    948.039702  ...   251.135000  4.023900e+05\n","Current Liabil     2003.0   2107.014083  ...  1074.699000  1.057180e+05\n","Total Liabil       2658.0  13558.881777  ...  5332.281000  2.426049e+06\n","NI                 2664.0    489.509873  ...   263.813000  5.525600e+04\n","PP&E               2534.0   2655.320498  ...  1090.268000  2.596510e+05\n","Retained Earnings  2656.0   2312.222038  ...   917.807500  4.018480e+05\n","Sales              2664.0   5907.023291  ...  3477.726000  5.214260e+05\n","Interest Expenses  2222.0    171.782209  ...   129.880000  1.653000e+04\n","GGROUP             2666.0   3482.070518  ...  4510.000000  6.010000e+03\n","Total Equity       2658.0   3830.273234  ...  2339.679750  2.648100e+05\n","\n","[19 rows x 8 columns]"]},"metadata":{"tags":[]},"execution_count":175}]},{"cell_type":"markdown","metadata":{"id":"RwKMHPWNdQnV"},"source":["## Step 2. Create new dataframe for financial ratios\n","\n","Create a list of financial ratio variables that would help explain the cross-sectional variation in returns.\n","\n","- Cash/Total Assets (Do firms with more cash at end of 2019 have\n","higher stock returns in 2020 after Covid shock?)\n","- Cash/Current_Liabilities (Do firms with more cash to cover current liabilities at end of 2019 have higher stock returns in 2020\n","after Covid shock?)\n","- Cash/Interest Expense (Do firms with more cash to cover\n","immediate interest expenses have higher stock returns in 2020\n","after Covid shock?)\n","- Quick Ratio (Do firms with more liquid assets to cover current\n","liabilities at end of 2019 have higher stock returns in 2020 after\n","Covid shock?)\n","- Long_Term_Debt/Total Assets (Do firms with lower Long Term\n","Debt at end of 2019 have higher stock returns in 2020 after Covid\n","shock?)\n","- EBIT/Interest Expense (Do firms with high 2019 earnings to cover\n","interest expenses have higher stock returns in 2020 after Covid\n","shock?)\n","- Fixed Assets/Total_Assets (Do firms with higher fixed assets\n","perform better or worse in 2020 after Covid shock?)\n","- Net Sales/# Employees = Sales per Employee (Do firms that have higher versus lower reliance on labor perform better or worse in 2020 after Covid shock?)\n","- Geographic Location (Do firms in certain States, Cities or ZipCodes have higher versus lower stock returns in 2020 after Covid-19 shock? Is there an economic story to support these patterns?)"]},{"cell_type":"code","metadata":{"colab":{"base_uri":"https://localhost:8080/","height":0},"id":"Dp43e8ZOdQne","executionInfo":{"status":"ok","timestamp":1625713107709,"user_tz":240,"elapsed":63,"user":{"displayName":"Tzu-Hua Huang","photoUrl":"https://lh3.googleusercontent.com/a-/AOh14GgbXOdZWwXGVTDuMHbeSme5C5oKAeF6I6ojY6qc=s64","userId":"01458545184256376706"}},"outputId":"84a393b8-b841-41c6-f52b-5c252906dd62"},"source":["df1 = pd.DataFrame()\n","df1['TICKER'] = df['TICKER']\n","df1['RetEarly2020'] = df['RetEarly2020']\n","df1['RetLate2020'] = df['RetLate2020']\n","\n","# Cash/Total Assets\n","df1['Cash/TA'] = df['Cash']/df['Total Assets']\n","\n","# Cash/Current_Liabilities\n","df1['Cash/CL'] = df['Cash']/df['Current Liabil']\n","\n","# Cash/Interest Expense\n","df1['Cash/Interest Expense'] = df['Cash']/df['Interest Expenses']\n","\n","# Quick ratio = (CA - Inventories) / CL\n","df1['Quick Ratio'] = (df['Current Assets'] - df['Inventories'])/df['Current Liabil']\n","\n","# Long_Term_Debt/Total Assets\n","df1['Long-term Debt/TA'] = df['Long-term Debt']/df['Total Assets']\n","\n","# EBIT/Interest Expense\n","df1['EBIT/Interest Expense'] = df['EBIT']/df['Interest Expenses']\n","\n","# Fixed Assets/Total_Assets\n","df1['Fixed Assets/TA'] = df['PP&E']/df['Total Assets']\n","\n","# Net Sales/# Employees = Sales per Employee\n","df1['Sales per Employee'] = df['Sales']/df['#Employees']\n","\n","df1.sample(5)"],"execution_count":176,"outputs":[{"output_type":"execute_result","data":{"text/html":["<div>\n","<style scoped>\n","    .dataframe tbody tr th:only-of-type {\n","        vertical-align: middle;\n","    }\n","\n","    .dataframe tbody tr th {\n","        vertical-align: top;\n","    }\n","\n","    .dataframe thead th {\n","        text-align: right;\n","    }\n","</style>\n","<table border=\"1\" class=\"dataframe\">\n","  <thead>\n","    <tr style=\"text-align: right;\">\n","      <th></th>\n","      <th>TICKER</th>\n","      <th>RetEarly2020</th>\n","      <th>RetLate2020</th>\n","      <th>Cash/TA</th>\n","      <th>Cash/CL</th>\n","      <th>Cash/Interest Expense</th>\n","      <th>Quick Ratio</th>\n","      <th>Long-term Debt/TA</th>\n","      <th>EBIT/Interest Expense</th>\n","      <th>Fixed Assets/TA</th>\n","      <th>Sales per Employee</th>\n","    </tr>\n","  </thead>\n","  <tbody>\n","    <tr>\n","      <th>668</th>\n","      <td>MEDP</td>\n","      <td>-0.127053</td>\n","      <td>0.896975</td>\n","      <td>0.115408</td>\n","      <td>0.384604</td>\n","      <td>112.945205</td>\n","      <td>NaN</td>\n","      <td>0.039553</td>\n","      <td>109.642979</td>\n","      <td>0.086997</td>\n","      <td>0.245991</td>\n","    </tr>\n","    <tr>\n","      <th>967</th>\n","      <td>DELL</td>\n","      <td>-0.230395</td>\n","      <td>0.853097</td>\n","      <td>0.084401</td>\n","      <td>0.191246</td>\n","      <td>3.750280</td>\n","      <td>0.640289</td>\n","      <td>0.384306</td>\n","      <td>1.318131</td>\n","      <td>0.065917</td>\n","      <td>0.558509</td>\n","    </tr>\n","    <tr>\n","      <th>2108</th>\n","      <td>HSII</td>\n","      <td>-0.303076</td>\n","      <td>0.331331</td>\n","      <td>0.394317</td>\n","      <td>0.963891</td>\n","      <td>NaN</td>\n","      <td>1.431862</td>\n","      <td>0.094042</td>\n","      <td>NaN</td>\n","      <td>0.151676</td>\n","      <td>0.407648</td>\n","    </tr>\n","    <tr>\n","      <th>151</th>\n","      <td>KSU</td>\n","      <td>-0.167015</td>\n","      <td>0.616383</td>\n","      <td>0.015204</td>\n","      <td>0.302870</td>\n","      <td>1.283865</td>\n","      <td>1.176471</td>\n","      <td>0.338589</td>\n","      <td>9.060397</td>\n","      <td>0.677361</td>\n","      <td>0.407102</td>\n","    </tr>\n","    <tr>\n","      <th>1892</th>\n","      <td>REV</td>\n","      <td>-0.489729</td>\n","      <td>0.086917</td>\n","      <td>0.034993</td>\n","      <td>0.108998</td>\n","      <td>0.493845</td>\n","      <td>0.692862</td>\n","      <td>1.006341</td>\n","      <td>0.386837</td>\n","      <td>0.170972</td>\n","      <td>0.342648</td>\n","    </tr>\n","  </tbody>\n","</table>\n","</div>"],"text/plain":["     TICKER  RetEarly2020  ...  Fixed Assets/TA  Sales per Employee\n","668    MEDP     -0.127053  ...         0.086997            0.245991\n","967    DELL     -0.230395  ...         0.065917            0.558509\n","2108   HSII     -0.303076  ...         0.151676            0.407648\n","151     KSU     -0.167015  ...         0.677361            0.407102\n","1892    REV     -0.489729  ...         0.170972            0.342648\n","\n","[5 rows x 11 columns]"]},"metadata":{"tags":[]},"execution_count":176}]},{"cell_type":"markdown","metadata":{"id":"yZqMfPXL47Su"},"source":["## Step 3. Check data "]},{"cell_type":"markdown","metadata":{"id":"M-zHH7vx37WW"},"source":["### 3.a Check for extreme/inf values"]},{"cell_type":"code","metadata":{"colab":{"base_uri":"https://localhost:8080/","height":359},"id":"OFBCszC0bJbF","executionInfo":{"status":"ok","timestamp":1625713107710,"user_tz":240,"elapsed":60,"user":{"displayName":"Tzu-Hua Huang","photoUrl":"https://lh3.googleusercontent.com/a-/AOh14GgbXOdZWwXGVTDuMHbeSme5C5oKAeF6I6ojY6qc=s64","userId":"01458545184256376706"}},"outputId":"ae999bb8-4465-4f23-ae05-1bc59196dd09"},"source":["df1.describe().T"],"execution_count":177,"outputs":[{"output_type":"execute_result","data":{"text/html":["<div>\n","<style scoped>\n","    .dataframe tbody tr th:only-of-type {\n","        vertical-align: middle;\n","    }\n","\n","    .dataframe tbody tr th {\n","        vertical-align: top;\n","    }\n","\n","    .dataframe thead th {\n","        text-align: right;\n","    }\n","</style>\n","<table border=\"1\" class=\"dataframe\">\n","  <thead>\n","    <tr style=\"text-align: right;\">\n","      <th></th>\n","      <th>count</th>\n","      <th>mean</th>\n","      <th>std</th>\n","      <th>min</th>\n","      <th>25%</th>\n","      <th>50%</th>\n","      <th>75%</th>\n","      <th>max</th>\n","    </tr>\n","  </thead>\n","  <tbody>\n","    <tr>\n","      <th>RetEarly2020</th>\n","      <td>2666.0</td>\n","      <td>-0.290915</td>\n","      <td>0.313196</td>\n","      <td>-0.908665</td>\n","      <td>-0.440401</td>\n","      <td>-0.312152</td>\n","      <td>-0.171201</td>\n","      <td>7.512067</td>\n","    </tr>\n","    <tr>\n","      <th>RetLate2020</th>\n","      <td>2666.0</td>\n","      <td>0.804400</td>\n","      <td>1.074286</td>\n","      <td>-0.858458</td>\n","      <td>0.252418</td>\n","      <td>0.547064</td>\n","      <td>1.000302</td>\n","      <td>18.993847</td>\n","    </tr>\n","    <tr>\n","      <th>Cash/TA</th>\n","      <td>2664.0</td>\n","      <td>0.187950</td>\n","      <td>0.244027</td>\n","      <td>0.000000</td>\n","      <td>0.027972</td>\n","      <td>0.076713</td>\n","      <td>0.233408</td>\n","      <td>0.997352</td>\n","    </tr>\n","    <tr>\n","      <th>Cash/CL</th>\n","      <td>2003.0</td>\n","      <td>2.083432</td>\n","      <td>5.662733</td>\n","      <td>0.000000</td>\n","      <td>0.206512</td>\n","      <td>0.579562</td>\n","      <td>1.635736</td>\n","      <td>131.641988</td>\n","    </tr>\n","    <tr>\n","      <th>Cash/Interest Expense</th>\n","      <td>2221.0</td>\n","      <td>inf</td>\n","      <td>NaN</td>\n","      <td>0.000000</td>\n","      <td>1.683858</td>\n","      <td>6.729831</td>\n","      <td>34.755518</td>\n","      <td>inf</td>\n","    </tr>\n","    <tr>\n","      <th>Quick Ratio</th>\n","      <td>1978.0</td>\n","      <td>2.855418</td>\n","      <td>5.816238</td>\n","      <td>0.056552</td>\n","      <td>0.906597</td>\n","      <td>1.434861</td>\n","      <td>2.648404</td>\n","      <td>132.760649</td>\n","    </tr>\n","    <tr>\n","      <th>Long-term Debt/TA</th>\n","      <td>2657.0</td>\n","      <td>0.262368</td>\n","      <td>0.240826</td>\n","      <td>0.000000</td>\n","      <td>0.053111</td>\n","      <td>0.232472</td>\n","      <td>0.402346</td>\n","      <td>3.092259</td>\n","    </tr>\n","    <tr>\n","      <th>EBIT/Interest Expense</th>\n","      <td>2222.0</td>\n","      <td>NaN</td>\n","      <td>NaN</td>\n","      <td>-inf</td>\n","      <td>0.934564</td>\n","      <td>3.432021</td>\n","      <td>9.559087</td>\n","      <td>inf</td>\n","    </tr>\n","    <tr>\n","      <th>Fixed Assets/TA</th>\n","      <td>2534.0</td>\n","      <td>0.212759</td>\n","      <td>0.236853</td>\n","      <td>0.000000</td>\n","      <td>0.033470</td>\n","      <td>0.120424</td>\n","      <td>0.299720</td>\n","      <td>0.954753</td>\n","    </tr>\n","    <tr>\n","      <th>Sales per Employee</th>\n","      <td>2647.0</td>\n","      <td>inf</td>\n","      <td>NaN</td>\n","      <td>0.000000</td>\n","      <td>0.222206</td>\n","      <td>0.353485</td>\n","      <td>0.706061</td>\n","      <td>inf</td>\n","    </tr>\n","  </tbody>\n","</table>\n","</div>"],"text/plain":["                        count      mean  ...        75%         max\n","RetEarly2020           2666.0 -0.290915  ...  -0.171201    7.512067\n","RetLate2020            2666.0  0.804400  ...   1.000302   18.993847\n","Cash/TA                2664.0  0.187950  ...   0.233408    0.997352\n","Cash/CL                2003.0  2.083432  ...   1.635736  131.641988\n","Cash/Interest Expense  2221.0       inf  ...  34.755518         inf\n","Quick Ratio            1978.0  2.855418  ...   2.648404  132.760649\n","Long-term Debt/TA      2657.0  0.262368  ...   0.402346    3.092259\n","EBIT/Interest Expense  2222.0       NaN  ...   9.559087         inf\n","Fixed Assets/TA        2534.0  0.212759  ...   0.299720    0.954753\n","Sales per Employee     2647.0       inf  ...   0.706061         inf\n","\n","[10 rows x 8 columns]"]},"metadata":{"tags":[]},"execution_count":177}]},{"cell_type":"markdown","metadata":{"id":"qj3RmIJQbJbF"},"source":["- From the previous descriptive statistics, we find several fields have minimum of 0. For those columns, if we use them as denominator for calculating financial ratios, we will receive infinity values as the above descriptive statistics table shows. Thus, we will invert the relevant ratios using these variables as denominator.\n","\n","- For `Cash/Interest Expense`, both of the variables have 0, if we consider the number of problematic values compared to the size of data set, we may not to consider this ratio in the model.\n","\n","- For `Sales per Employee`, similarly, both of the variables have 0. Compared to the size of data set, the number of extreme and missing values are acceptable. Rather than invert the ratio, we can delete the rows with extreme and missing values."]},{"cell_type":"code","metadata":{"id":"WiLTrnuQ638-","colab":{"base_uri":"https://localhost:8080/","height":359},"executionInfo":{"status":"ok","timestamp":1625713107710,"user_tz":240,"elapsed":56,"user":{"displayName":"Tzu-Hua Huang","photoUrl":"https://lh3.googleusercontent.com/a-/AOh14GgbXOdZWwXGVTDuMHbeSme5C5oKAeF6I6ojY6qc=s64","userId":"01458545184256376706"}},"outputId":"283b1237-3568-4757-e3a5-9c4731a5c98b"},"source":["# df1.drop(columns=['Cash/Interest Expense'], inplace=True)\n","\n","# invert ratios with infinity values\n","# drop the columns first\n","df1.drop(columns=['EBIT/Interest Expense'], inplace=True)\n","\n","# invert EBIT/Interest Expense to Interest Expense/EBIT\n","df1['Interest Expense/EBIT'] = df['Interest Expenses']/df['EBIT']\n","df1.describe().T"],"execution_count":178,"outputs":[{"output_type":"execute_result","data":{"text/html":["<div>\n","<style scoped>\n","    .dataframe tbody tr th:only-of-type {\n","        vertical-align: middle;\n","    }\n","\n","    .dataframe tbody tr th {\n","        vertical-align: top;\n","    }\n","\n","    .dataframe thead th {\n","        text-align: right;\n","    }\n","</style>\n","<table border=\"1\" class=\"dataframe\">\n","  <thead>\n","    <tr style=\"text-align: right;\">\n","      <th></th>\n","      <th>count</th>\n","      <th>mean</th>\n","      <th>std</th>\n","      <th>min</th>\n","      <th>25%</th>\n","      <th>50%</th>\n","      <th>75%</th>\n","      <th>max</th>\n","    </tr>\n","  </thead>\n","  <tbody>\n","    <tr>\n","      <th>RetEarly2020</th>\n","      <td>2666.0</td>\n","      <td>-0.290915</td>\n","      <td>0.313196</td>\n","      <td>-0.908665</td>\n","      <td>-0.440401</td>\n","      <td>-0.312152</td>\n","      <td>-0.171201</td>\n","      <td>7.512067</td>\n","    </tr>\n","    <tr>\n","      <th>RetLate2020</th>\n","      <td>2666.0</td>\n","      <td>0.804400</td>\n","      <td>1.074286</td>\n","      <td>-0.858458</td>\n","      <td>0.252418</td>\n","      <td>0.547064</td>\n","      <td>1.000302</td>\n","      <td>18.993847</td>\n","    </tr>\n","    <tr>\n","      <th>Cash/TA</th>\n","      <td>2664.0</td>\n","      <td>0.187950</td>\n","      <td>0.244027</td>\n","      <td>0.000000</td>\n","      <td>0.027972</td>\n","      <td>0.076713</td>\n","      <td>0.233408</td>\n","      <td>0.997352</td>\n","    </tr>\n","    <tr>\n","      <th>Cash/CL</th>\n","      <td>2003.0</td>\n","      <td>2.083432</td>\n","      <td>5.662733</td>\n","      <td>0.000000</td>\n","      <td>0.206512</td>\n","      <td>0.579562</td>\n","      <td>1.635736</td>\n","      <td>131.641988</td>\n","    </tr>\n","    <tr>\n","      <th>Cash/Interest Expense</th>\n","      <td>2221.0</td>\n","      <td>inf</td>\n","      <td>NaN</td>\n","      <td>0.000000</td>\n","      <td>1.683858</td>\n","      <td>6.729831</td>\n","      <td>34.755518</td>\n","      <td>inf</td>\n","    </tr>\n","    <tr>\n","      <th>Quick Ratio</th>\n","      <td>1978.0</td>\n","      <td>2.855418</td>\n","      <td>5.816238</td>\n","      <td>0.056552</td>\n","      <td>0.906597</td>\n","      <td>1.434861</td>\n","      <td>2.648404</td>\n","      <td>132.760649</td>\n","    </tr>\n","    <tr>\n","      <th>Long-term Debt/TA</th>\n","      <td>2657.0</td>\n","      <td>0.262368</td>\n","      <td>0.240826</td>\n","      <td>0.000000</td>\n","      <td>0.053111</td>\n","      <td>0.232472</td>\n","      <td>0.402346</td>\n","      <td>3.092259</td>\n","    </tr>\n","    <tr>\n","      <th>Fixed Assets/TA</th>\n","      <td>2534.0</td>\n","      <td>0.212759</td>\n","      <td>0.236853</td>\n","      <td>0.000000</td>\n","      <td>0.033470</td>\n","      <td>0.120424</td>\n","      <td>0.299720</td>\n","      <td>0.954753</td>\n","    </tr>\n","    <tr>\n","      <th>Sales per Employee</th>\n","      <td>2647.0</td>\n","      <td>inf</td>\n","      <td>NaN</td>\n","      <td>0.000000</td>\n","      <td>0.222206</td>\n","      <td>0.353485</td>\n","      <td>0.706061</td>\n","      <td>inf</td>\n","    </tr>\n","    <tr>\n","      <th>Interest Expense/EBIT</th>\n","      <td>2222.0</td>\n","      <td>0.147208</td>\n","      <td>2.309836</td>\n","      <td>-62.025641</td>\n","      <td>0.003666</td>\n","      <td>0.132971</td>\n","      <td>0.350891</td>\n","      <td>15.335968</td>\n","    </tr>\n","  </tbody>\n","</table>\n","</div>"],"text/plain":["                        count      mean  ...        75%         max\n","RetEarly2020           2666.0 -0.290915  ...  -0.171201    7.512067\n","RetLate2020            2666.0  0.804400  ...   1.000302   18.993847\n","Cash/TA                2664.0  0.187950  ...   0.233408    0.997352\n","Cash/CL                2003.0  2.083432  ...   1.635736  131.641988\n","Cash/Interest Expense  2221.0       inf  ...  34.755518         inf\n","Quick Ratio            1978.0  2.855418  ...   2.648404  132.760649\n","Long-term Debt/TA      2657.0  0.262368  ...   0.402346    3.092259\n","Fixed Assets/TA        2534.0  0.212759  ...   0.299720    0.954753\n","Sales per Employee     2647.0       inf  ...   0.706061         inf\n","Interest Expense/EBIT  2222.0  0.147208  ...   0.350891   15.335968\n","\n","[10 rows x 8 columns]"]},"metadata":{"tags":[]},"execution_count":178}]},{"cell_type":"markdown","metadata":{"id":"csXE2mqJ7G2N"},"source":["### 3.b Check for missing values"]},{"cell_type":"code","metadata":{"colab":{"base_uri":"https://localhost:8080/"},"id":"7JJ3J763vt7g","executionInfo":{"status":"ok","timestamp":1625713107711,"user_tz":240,"elapsed":52,"user":{"displayName":"Tzu-Hua Huang","photoUrl":"https://lh3.googleusercontent.com/a-/AOh14GgbXOdZWwXGVTDuMHbeSme5C5oKAeF6I6ojY6qc=s64","userId":"01458545184256376706"}},"outputId":"f269d41b-7df6-4e5b-98a4-3c6fccb2dcbb"},"source":["# check columns with na values\n","df1.isna().sum()"],"execution_count":179,"outputs":[{"output_type":"execute_result","data":{"text/plain":["TICKER                     0\n","RetEarly2020               0\n","RetLate2020                0\n","Cash/TA                    2\n","Cash/CL                  663\n","Cash/Interest Expense    445\n","Quick Ratio              688\n","Long-term Debt/TA          9\n","Fixed Assets/TA          132\n","Sales per Employee        19\n","Interest Expense/EBIT    444\n","dtype: int64"]},"metadata":{"tags":[]},"execution_count":179}]},{"cell_type":"markdown","metadata":{"id":"YyuSiyKBwttc"},"source":["- All of the ratios have NaN values, compared to the size of the whole data set, rather than delete them directly, we replace the missing values with the sample mean."]},{"cell_type":"code","metadata":{"colab":{"base_uri":"https://localhost:8080/","height":359},"id":"Gs62GHRc7lAS","executionInfo":{"status":"ok","timestamp":1625713108113,"user_tz":240,"elapsed":440,"user":{"displayName":"Tzu-Hua Huang","photoUrl":"https://lh3.googleusercontent.com/a-/AOh14GgbXOdZWwXGVTDuMHbeSme5C5oKAeF6I6ojY6qc=s64","userId":"01458545184256376706"}},"outputId":"6f81b8dc-9104-4066-a95e-71cc158eda67"},"source":["df1.describe().T"],"execution_count":180,"outputs":[{"output_type":"execute_result","data":{"text/html":["<div>\n","<style scoped>\n","    .dataframe tbody tr th:only-of-type {\n","        vertical-align: middle;\n","    }\n","\n","    .dataframe tbody tr th {\n","        vertical-align: top;\n","    }\n","\n","    .dataframe thead th {\n","        text-align: right;\n","    }\n","</style>\n","<table border=\"1\" class=\"dataframe\">\n","  <thead>\n","    <tr style=\"text-align: right;\">\n","      <th></th>\n","      <th>count</th>\n","      <th>mean</th>\n","      <th>std</th>\n","      <th>min</th>\n","      <th>25%</th>\n","      <th>50%</th>\n","      <th>75%</th>\n","      <th>max</th>\n","    </tr>\n","  </thead>\n","  <tbody>\n","    <tr>\n","      <th>RetEarly2020</th>\n","      <td>2666.0</td>\n","      <td>-0.290915</td>\n","      <td>0.313196</td>\n","      <td>-0.908665</td>\n","      <td>-0.440401</td>\n","      <td>-0.312152</td>\n","      <td>-0.171201</td>\n","      <td>7.512067</td>\n","    </tr>\n","    <tr>\n","      <th>RetLate2020</th>\n","      <td>2666.0</td>\n","      <td>0.804400</td>\n","      <td>1.074286</td>\n","      <td>-0.858458</td>\n","      <td>0.252418</td>\n","      <td>0.547064</td>\n","      <td>1.000302</td>\n","      <td>18.993847</td>\n","    </tr>\n","    <tr>\n","      <th>Cash/TA</th>\n","      <td>2664.0</td>\n","      <td>0.187950</td>\n","      <td>0.244027</td>\n","      <td>0.000000</td>\n","      <td>0.027972</td>\n","      <td>0.076713</td>\n","      <td>0.233408</td>\n","      <td>0.997352</td>\n","    </tr>\n","    <tr>\n","      <th>Cash/CL</th>\n","      <td>2003.0</td>\n","      <td>2.083432</td>\n","      <td>5.662733</td>\n","      <td>0.000000</td>\n","      <td>0.206512</td>\n","      <td>0.579562</td>\n","      <td>1.635736</td>\n","      <td>131.641988</td>\n","    </tr>\n","    <tr>\n","      <th>Cash/Interest Expense</th>\n","      <td>2221.0</td>\n","      <td>inf</td>\n","      <td>NaN</td>\n","      <td>0.000000</td>\n","      <td>1.683858</td>\n","      <td>6.729831</td>\n","      <td>34.755518</td>\n","      <td>inf</td>\n","    </tr>\n","    <tr>\n","      <th>Quick Ratio</th>\n","      <td>1978.0</td>\n","      <td>2.855418</td>\n","      <td>5.816238</td>\n","      <td>0.056552</td>\n","      <td>0.906597</td>\n","      <td>1.434861</td>\n","      <td>2.648404</td>\n","      <td>132.760649</td>\n","    </tr>\n","    <tr>\n","      <th>Long-term Debt/TA</th>\n","      <td>2657.0</td>\n","      <td>0.262368</td>\n","      <td>0.240826</td>\n","      <td>0.000000</td>\n","      <td>0.053111</td>\n","      <td>0.232472</td>\n","      <td>0.402346</td>\n","      <td>3.092259</td>\n","    </tr>\n","    <tr>\n","      <th>Fixed Assets/TA</th>\n","      <td>2534.0</td>\n","      <td>0.212759</td>\n","      <td>0.236853</td>\n","      <td>0.000000</td>\n","      <td>0.033470</td>\n","      <td>0.120424</td>\n","      <td>0.299720</td>\n","      <td>0.954753</td>\n","    </tr>\n","    <tr>\n","      <th>Sales per Employee</th>\n","      <td>2647.0</td>\n","      <td>inf</td>\n","      <td>NaN</td>\n","      <td>0.000000</td>\n","      <td>0.222206</td>\n","      <td>0.353485</td>\n","      <td>0.706061</td>\n","      <td>inf</td>\n","    </tr>\n","    <tr>\n","      <th>Interest Expense/EBIT</th>\n","      <td>2222.0</td>\n","      <td>0.147208</td>\n","      <td>2.309836</td>\n","      <td>-62.025641</td>\n","      <td>0.003666</td>\n","      <td>0.132971</td>\n","      <td>0.350891</td>\n","      <td>15.335968</td>\n","    </tr>\n","  </tbody>\n","</table>\n","</div>"],"text/plain":["                        count      mean  ...        75%         max\n","RetEarly2020           2666.0 -0.290915  ...  -0.171201    7.512067\n","RetLate2020            2666.0  0.804400  ...   1.000302   18.993847\n","Cash/TA                2664.0  0.187950  ...   0.233408    0.997352\n","Cash/CL                2003.0  2.083432  ...   1.635736  131.641988\n","Cash/Interest Expense  2221.0       inf  ...  34.755518         inf\n","Quick Ratio            1978.0  2.855418  ...   2.648404  132.760649\n","Long-term Debt/TA      2657.0  0.262368  ...   0.402346    3.092259\n","Fixed Assets/TA        2534.0  0.212759  ...   0.299720    0.954753\n","Sales per Employee     2647.0       inf  ...   0.706061         inf\n","Interest Expense/EBIT  2222.0  0.147208  ...   0.350891   15.335968\n","\n","[10 rows x 8 columns]"]},"metadata":{"tags":[]},"execution_count":180}]},{"cell_type":"code","metadata":{"colab":{"base_uri":"https://localhost:8080/","height":359},"id":"tVfPeJIf7geu","executionInfo":{"status":"ok","timestamp":1625713108114,"user_tz":240,"elapsed":172,"user":{"displayName":"Tzu-Hua Huang","photoUrl":"https://lh3.googleusercontent.com/a-/AOh14GgbXOdZWwXGVTDuMHbeSme5C5oKAeF6I6ojY6qc=s64","userId":"01458545184256376706"}},"outputId":"065875eb-b655-477d-c9e7-976b813bd9f0"},"source":["df1['Cash/TA'].fillna((df1['Cash/TA'].mean()), inplace=True)\n","df1['Cash/CL'].fillna((df1['Cash/CL'].mean()), inplace=True)\n","df1['Cash/Interest Expense'].fillna((df1['Cash/Interest Expense'].mean()), inplace=True)\n","df1['Quick Ratio'].fillna((df1['Quick Ratio'].mean()), inplace=True)\n","df1['Long-term Debt/TA'].fillna((df1['Long-term Debt/TA'].mean()), inplace=True)\n","df1['Fixed Assets/TA'].fillna((df1['Fixed Assets/TA'].mean()), inplace=True)\n","df1['Sales per Employee'].fillna((df1['Sales per Employee'].mean()), inplace=True)\n","df1['Interest Expense/EBIT'].fillna((df1['Interest Expense/EBIT'].mean()), inplace=True)\n","df1.describe().T"],"execution_count":181,"outputs":[{"output_type":"execute_result","data":{"text/html":["<div>\n","<style scoped>\n","    .dataframe tbody tr th:only-of-type {\n","        vertical-align: middle;\n","    }\n","\n","    .dataframe tbody tr th {\n","        vertical-align: top;\n","    }\n","\n","    .dataframe thead th {\n","        text-align: right;\n","    }\n","</style>\n","<table border=\"1\" class=\"dataframe\">\n","  <thead>\n","    <tr style=\"text-align: right;\">\n","      <th></th>\n","      <th>count</th>\n","      <th>mean</th>\n","      <th>std</th>\n","      <th>min</th>\n","      <th>25%</th>\n","      <th>50%</th>\n","      <th>75%</th>\n","      <th>max</th>\n","    </tr>\n","  </thead>\n","  <tbody>\n","    <tr>\n","      <th>RetEarly2020</th>\n","      <td>2666.0</td>\n","      <td>-0.290915</td>\n","      <td>0.313196</td>\n","      <td>-0.908665</td>\n","      <td>-0.440401</td>\n","      <td>-0.312152</td>\n","      <td>-0.171201</td>\n","      <td>7.512067</td>\n","    </tr>\n","    <tr>\n","      <th>RetLate2020</th>\n","      <td>2666.0</td>\n","      <td>0.804400</td>\n","      <td>1.074286</td>\n","      <td>-0.858458</td>\n","      <td>0.252418</td>\n","      <td>0.547064</td>\n","      <td>1.000302</td>\n","      <td>18.993847</td>\n","    </tr>\n","    <tr>\n","      <th>Cash/TA</th>\n","      <td>2666.0</td>\n","      <td>0.187950</td>\n","      <td>0.243935</td>\n","      <td>0.000000</td>\n","      <td>0.027986</td>\n","      <td>0.076742</td>\n","      <td>0.233232</td>\n","      <td>0.997352</td>\n","    </tr>\n","    <tr>\n","      <th>Cash/CL</th>\n","      <td>2666.0</td>\n","      <td>2.083432</td>\n","      <td>4.908056</td>\n","      <td>0.000000</td>\n","      <td>0.288087</td>\n","      <td>1.087557</td>\n","      <td>2.083432</td>\n","      <td>131.641988</td>\n","    </tr>\n","    <tr>\n","      <th>Cash/Interest Expense</th>\n","      <td>2666.0</td>\n","      <td>inf</td>\n","      <td>NaN</td>\n","      <td>0.000000</td>\n","      <td>2.261563</td>\n","      <td>11.883578</td>\n","      <td>991.921245</td>\n","      <td>inf</td>\n","    </tr>\n","    <tr>\n","      <th>Quick Ratio</th>\n","      <td>2666.0</td>\n","      <td>2.855418</td>\n","      <td>5.009529</td>\n","      <td>0.056552</td>\n","      <td>1.080186</td>\n","      <td>2.020087</td>\n","      <td>2.855418</td>\n","      <td>132.760649</td>\n","    </tr>\n","    <tr>\n","      <th>Long-term Debt/TA</th>\n","      <td>2666.0</td>\n","      <td>0.262368</td>\n","      <td>0.240419</td>\n","      <td>0.000000</td>\n","      <td>0.053332</td>\n","      <td>0.233069</td>\n","      <td>0.401910</td>\n","      <td>3.092259</td>\n","    </tr>\n","    <tr>\n","      <th>Fixed Assets/TA</th>\n","      <td>2666.0</td>\n","      <td>0.212759</td>\n","      <td>0.230913</td>\n","      <td>0.000000</td>\n","      <td>0.039648</td>\n","      <td>0.132109</td>\n","      <td>0.281346</td>\n","      <td>0.954753</td>\n","    </tr>\n","    <tr>\n","      <th>Sales per Employee</th>\n","      <td>2666.0</td>\n","      <td>inf</td>\n","      <td>NaN</td>\n","      <td>0.000000</td>\n","      <td>0.223460</td>\n","      <td>0.356995</td>\n","      <td>0.722607</td>\n","      <td>inf</td>\n","    </tr>\n","    <tr>\n","      <th>Interest Expense/EBIT</th>\n","      <td>2666.0</td>\n","      <td>0.147208</td>\n","      <td>2.108661</td>\n","      <td>-62.025641</td>\n","      <td>0.030440</td>\n","      <td>0.147208</td>\n","      <td>0.288160</td>\n","      <td>15.335968</td>\n","    </tr>\n","  </tbody>\n","</table>\n","</div>"],"text/plain":["                        count      mean  ...         75%         max\n","RetEarly2020           2666.0 -0.290915  ...   -0.171201    7.512067\n","RetLate2020            2666.0  0.804400  ...    1.000302   18.993847\n","Cash/TA                2666.0  0.187950  ...    0.233232    0.997352\n","Cash/CL                2666.0  2.083432  ...    2.083432  131.641988\n","Cash/Interest Expense  2666.0       inf  ...  991.921245         inf\n","Quick Ratio            2666.0  2.855418  ...    2.855418  132.760649\n","Long-term Debt/TA      2666.0  0.262368  ...    0.401910    3.092259\n","Fixed Assets/TA        2666.0  0.212759  ...    0.281346    0.954753\n","Sales per Employee     2666.0       inf  ...    0.722607         inf\n","Interest Expense/EBIT  2666.0  0.147208  ...    0.288160   15.335968\n","\n","[10 rows x 8 columns]"]},"metadata":{"tags":[]},"execution_count":181}]},{"cell_type":"markdown","metadata":{"id":"GgtXlxA90Bjn"},"source":["Now, we can drop the observations with infinity values."]},{"cell_type":"code","metadata":{"colab":{"base_uri":"https://localhost:8080/","height":359},"id":"HOKHj-oowZrX","executionInfo":{"status":"ok","timestamp":1625713108115,"user_tz":240,"elapsed":159,"user":{"displayName":"Tzu-Hua Huang","photoUrl":"https://lh3.googleusercontent.com/a-/AOh14GgbXOdZWwXGVTDuMHbeSme5C5oKAeF6I6ojY6qc=s64","userId":"01458545184256376706"}},"outputId":"3255521a-d618-4df4-988b-cc00ad828d89"},"source":["df1 = df1.replace([np.inf, -np.inf], np.nan).dropna(axis=0)\n","df1.describe().T"],"execution_count":182,"outputs":[{"output_type":"execute_result","data":{"text/html":["<div>\n","<style scoped>\n","    .dataframe tbody tr th:only-of-type {\n","        vertical-align: middle;\n","    }\n","\n","    .dataframe tbody tr th {\n","        vertical-align: top;\n","    }\n","\n","    .dataframe thead th {\n","        text-align: right;\n","    }\n","</style>\n","<table border=\"1\" class=\"dataframe\">\n","  <thead>\n","    <tr style=\"text-align: right;\">\n","      <th></th>\n","      <th>count</th>\n","      <th>mean</th>\n","      <th>std</th>\n","      <th>min</th>\n","      <th>25%</th>\n","      <th>50%</th>\n","      <th>75%</th>\n","      <th>max</th>\n","    </tr>\n","  </thead>\n","  <tbody>\n","    <tr>\n","      <th>RetEarly2020</th>\n","      <td>2022.0</td>\n","      <td>-0.295033</td>\n","      <td>0.328269</td>\n","      <td>-0.908665</td>\n","      <td>-0.451292</td>\n","      <td>-0.311070</td>\n","      <td>-0.167882</td>\n","      <td>7.512067</td>\n","    </tr>\n","    <tr>\n","      <th>RetLate2020</th>\n","      <td>2022.0</td>\n","      <td>0.850017</td>\n","      <td>1.093635</td>\n","      <td>-0.761450</td>\n","      <td>0.291102</td>\n","      <td>0.592884</td>\n","      <td>1.063381</td>\n","      <td>18.993847</td>\n","    </tr>\n","    <tr>\n","      <th>Cash/TA</th>\n","      <td>2022.0</td>\n","      <td>0.162702</td>\n","      <td>0.213549</td>\n","      <td>0.000000</td>\n","      <td>0.026525</td>\n","      <td>0.074498</td>\n","      <td>0.193506</td>\n","      <td>0.997352</td>\n","    </tr>\n","    <tr>\n","      <th>Cash/CL</th>\n","      <td>2022.0</td>\n","      <td>1.566971</td>\n","      <td>4.662722</td>\n","      <td>0.000000</td>\n","      <td>0.221913</td>\n","      <td>0.642435</td>\n","      <td>2.083432</td>\n","      <td>131.641988</td>\n","    </tr>\n","    <tr>\n","      <th>Cash/Interest Expense</th>\n","      <td>2022.0</td>\n","      <td>493.038341</td>\n","      <td>6335.332240</td>\n","      <td>0.000000</td>\n","      <td>1.519326</td>\n","      <td>5.809825</td>\n","      <td>22.681178</td>\n","      <td>143093.000000</td>\n","    </tr>\n","    <tr>\n","      <th>Quick Ratio</th>\n","      <td>2022.0</td>\n","      <td>2.351488</td>\n","      <td>4.807304</td>\n","      <td>0.056552</td>\n","      <td>0.941378</td>\n","      <td>1.549811</td>\n","      <td>2.855418</td>\n","      <td>132.760649</td>\n","    </tr>\n","    <tr>\n","      <th>Long-term Debt/TA</th>\n","      <td>2022.0</td>\n","      <td>0.321280</td>\n","      <td>0.237723</td>\n","      <td>0.000000</td>\n","      <td>0.153549</td>\n","      <td>0.303907</td>\n","      <td>0.435809</td>\n","      <td>3.092259</td>\n","    </tr>\n","    <tr>\n","      <th>Fixed Assets/TA</th>\n","      <td>2022.0</td>\n","      <td>0.253329</td>\n","      <td>0.239436</td>\n","      <td>0.000000</td>\n","      <td>0.077808</td>\n","      <td>0.174078</td>\n","      <td>0.353765</td>\n","      <td>0.954753</td>\n","    </tr>\n","    <tr>\n","      <th>Sales per Employee</th>\n","      <td>2022.0</td>\n","      <td>1.903525</td>\n","      <td>20.402755</td>\n","      <td>0.000000</td>\n","      <td>0.224528</td>\n","      <td>0.379812</td>\n","      <td>0.801262</td>\n","      <td>786.261000</td>\n","    </tr>\n","    <tr>\n","      <th>Interest Expense/EBIT</th>\n","      <td>2022.0</td>\n","      <td>0.158576</td>\n","      <td>2.413454</td>\n","      <td>-62.025641</td>\n","      <td>0.031697</td>\n","      <td>0.151023</td>\n","      <td>0.364305</td>\n","      <td>15.335968</td>\n","    </tr>\n","  </tbody>\n","</table>\n","</div>"],"text/plain":["                        count        mean  ...        75%            max\n","RetEarly2020           2022.0   -0.295033  ...  -0.167882       7.512067\n","RetLate2020            2022.0    0.850017  ...   1.063381      18.993847\n","Cash/TA                2022.0    0.162702  ...   0.193506       0.997352\n","Cash/CL                2022.0    1.566971  ...   2.083432     131.641988\n","Cash/Interest Expense  2022.0  493.038341  ...  22.681178  143093.000000\n","Quick Ratio            2022.0    2.351488  ...   2.855418     132.760649\n","Long-term Debt/TA      2022.0    0.321280  ...   0.435809       3.092259\n","Fixed Assets/TA        2022.0    0.253329  ...   0.353765       0.954753\n","Sales per Employee     2022.0    1.903525  ...   0.801262     786.261000\n","Interest Expense/EBIT  2022.0    0.158576  ...   0.364305      15.335968\n","\n","[10 rows x 8 columns]"]},"metadata":{"tags":[]},"execution_count":182}]},{"cell_type":"markdown","metadata":{"id":"SoVQfWHHyWgL"},"source":["We have 2022 observations left. "]},{"cell_type":"markdown","metadata":{"id":"L4DGnbf6fsYj"},"source":["### 3.c Check for outliers - winsozrization\n","Winsorize each financial ratios by replacing 1 to 5% of the lowest/highest values."]},{"cell_type":"code","metadata":{"colab":{"base_uri":"https://localhost:8080/"},"id":"9EeJd2xNfwxx","executionInfo":{"status":"ok","timestamp":1625713108116,"user_tz":240,"elapsed":157,"user":{"displayName":"Tzu-Hua Huang","photoUrl":"https://lh3.googleusercontent.com/a-/AOh14GgbXOdZWwXGVTDuMHbeSme5C5oKAeF6I6ojY6qc=s64","userId":"01458545184256376706"}},"outputId":"a8290f37-7a83-43d0-b56c-961718a5f9c9"},"source":["df1.columns"],"execution_count":183,"outputs":[{"output_type":"execute_result","data":{"text/plain":["Index(['TICKER', 'RetEarly2020', 'RetLate2020', 'Cash/TA', 'Cash/CL',\n","       'Cash/Interest Expense', 'Quick Ratio', 'Long-term Debt/TA',\n","       'Fixed Assets/TA', 'Sales per Employee', 'Interest Expense/EBIT'],\n","      dtype='object')"]},"metadata":{"tags":[]},"execution_count":183}]},{"cell_type":"markdown","metadata":{"id":"58hIyTEKfsYk"},"source":["##### Cash/TA"]},{"cell_type":"code","metadata":{"colab":{"base_uri":"https://localhost:8080/","height":333},"id":"0_phSumZfsYk","executionInfo":{"status":"ok","timestamp":1625713108627,"user_tz":240,"elapsed":553,"user":{"displayName":"Tzu-Hua Huang","photoUrl":"https://lh3.googleusercontent.com/a-/AOh14GgbXOdZWwXGVTDuMHbeSme5C5oKAeF6I6ojY6qc=s64","userId":"01458545184256376706"}},"outputId":"91c61564-d46a-45dd-89fe-4c27a7ec97a5"},"source":["import matplotlib.pyplot as plt\n","import seaborn as sns\n","\n","# check outliers with boxplot\n","sns.boxplot(df1['Cash/TA'])\n","plt.show()"],"execution_count":184,"outputs":[{"output_type":"stream","text":["/usr/local/lib/python3.7/dist-packages/seaborn/_decorators.py:43: FutureWarning: Pass the following variable as a keyword arg: x. From version 0.12, the only valid positional argument will be `data`, and passing other arguments without an explicit keyword will result in an error or misinterpretation.\n","  FutureWarning\n"],"name":"stderr"},{"output_type":"display_data","data":{"image/png":"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\n","text/plain":["<Figure size 432x288 with 1 Axes>"]},"metadata":{"tags":[],"needs_background":"light"}}]},{"cell_type":"code","metadata":{"colab":{"base_uri":"https://localhost:8080/"},"id":"TL-S0oYafsYk","executionInfo":{"status":"ok","timestamp":1625713108638,"user_tz":240,"elapsed":337,"user":{"displayName":"Tzu-Hua Huang","photoUrl":"https://lh3.googleusercontent.com/a-/AOh14GgbXOdZWwXGVTDuMHbeSme5C5oKAeF6I6ojY6qc=s64","userId":"01458545184256376706"}},"outputId":"cb33aaa4-d7d7-459b-84c0-b1eb5d57e149"},"source":["# deal with outliers by winzoring\n","print(df1['Cash/TA'].quantile(0.00))\n","print(df1['Cash/TA'].quantile(0.90))"],"execution_count":185,"outputs":[{"output_type":"stream","text":["0.0\n","0.4839434290734715\n"],"name":"stdout"}]},{"cell_type":"code","metadata":{"colab":{"base_uri":"https://localhost:8080/"},"id":"XV-0-4OVy2P1","executionInfo":{"status":"ok","timestamp":1625713108638,"user_tz":240,"elapsed":213,"user":{"displayName":"Tzu-Hua Huang","photoUrl":"https://lh3.googleusercontent.com/a-/AOh14GgbXOdZWwXGVTDuMHbeSme5C5oKAeF6I6ojY6qc=s64","userId":"01458545184256376706"}},"outputId":"f5c461bf-9e94-4335-8f25-43ab2054d314"},"source":["# remove the effect of outliers\n","# according to the boxplot, replace 1% of the lowest and 2% the highest values by winsorizing\n","df1['Cash/TA'].clip(lower=df1['Cash/TA'].quantile(0.00),\n","                    upper=df1['Cash/TA'].quantile(0.90),\n","                    axis=0, inplace=True)\n","df1['Cash/TA'].describe()"],"execution_count":186,"outputs":[{"output_type":"execute_result","data":{"text/plain":["count    2022.000000\n","mean        0.140214\n","std         0.153353\n","min         0.000000\n","25%         0.026525\n","50%         0.074498\n","75%         0.193506\n","max         0.483943\n","Name: Cash/TA, dtype: float64"]},"metadata":{"tags":[]},"execution_count":186}]},{"cell_type":"code","metadata":{"colab":{"base_uri":"https://localhost:8080/","height":333},"id":"KVbtlyt8fsYl","executionInfo":{"status":"ok","timestamp":1625713108639,"user_tz":240,"elapsed":70,"user":{"displayName":"Tzu-Hua Huang","photoUrl":"https://lh3.googleusercontent.com/a-/AOh14GgbXOdZWwXGVTDuMHbeSme5C5oKAeF6I6ojY6qc=s64","userId":"01458545184256376706"}},"outputId":"88c1e294-25e6-4112-afa1-e78d9e8e716d"},"source":["# check if outliers exist after winzoring\n","sns.boxplot(df1['Cash/TA'])\n","plt.show()"],"execution_count":187,"outputs":[{"output_type":"stream","text":["/usr/local/lib/python3.7/dist-packages/seaborn/_decorators.py:43: FutureWarning: Pass the following variable as a keyword arg: x. From version 0.12, the only valid positional argument will be `data`, and passing other arguments without an explicit keyword will result in an error or misinterpretation.\n","  FutureWarning\n"],"name":"stderr"},{"output_type":"display_data","data":{"image/png":"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\n","text/plain":["<Figure size 432x288 with 1 Axes>"]},"metadata":{"tags":[],"needs_background":"light"}}]},{"cell_type":"markdown","metadata":{"id":"YVifKRdYfsYl"},"source":["##### Cash/CL"]},{"cell_type":"code","metadata":{"colab":{"base_uri":"https://localhost:8080/","height":333},"id":"CfnCfDKZfsYm","executionInfo":{"status":"ok","timestamp":1625713109314,"user_tz":240,"elapsed":729,"user":{"displayName":"Tzu-Hua Huang","photoUrl":"https://lh3.googleusercontent.com/a-/AOh14GgbXOdZWwXGVTDuMHbeSme5C5oKAeF6I6ojY6qc=s64","userId":"01458545184256376706"}},"outputId":"0ff34def-9e79-4d01-926c-bbf6bbefffc8"},"source":["# check outliers with boxplot\n","sns.boxplot(df1['Cash/CL'])\n","plt.show()"],"execution_count":188,"outputs":[{"output_type":"stream","text":["/usr/local/lib/python3.7/dist-packages/seaborn/_decorators.py:43: FutureWarning: Pass the following variable as a keyword arg: x. From version 0.12, the only valid positional argument will be `data`, and passing other arguments without an explicit keyword will result in an error or misinterpretation.\n","  FutureWarning\n"],"name":"stderr"},{"output_type":"display_data","data":{"image/png":"iVBORw0KGgoAAAANSUhEUgAAAWAAAAEGCAYAAABbzE8LAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADh0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uMy4yLjIsIGh0dHA6Ly9tYXRwbG90bGliLm9yZy+WH4yJAAAPq0lEQVR4nO3df2zc9X3H8dcbO4FAtpE6UbaaKqYyK2KstGBNjNEpdAEcJwqrtGlMmexpFdFgxNmyaSoiWkCyNlXd2CDaqErb9bJFpSprt5A52RLasWlS6WzaYEhCcqzuwIOSHCldfsfOe398v3bP9vecs+O799fm+ZBO9vdzn+/3877P+V53/tz5a3N3AQDq77LoAgDgvYoABoAgBDAABCGAASAIAQwAQRqn03np0qXe0tJSo1IAYH7q7+8/5u7LJrZPK4BbWlrU19c3e1UBwHuAmX0/q50lCAAIQgADQBACGACCEMAAEIQABoAgBDAABCGAASAIAQwAQQhgAAhCAANAEAIYAIIQwAAQhAAGgCAEMAAEIYABIAgBDABBCGAACEIAA0AQAhgAgkzrf8LN1LZt2/T8889Lkpqbm8faW1tbtXHjxnqUAAC5U5cALhaLOnqsJDU06q2zyZANp96px9AAkFt1CWBJUkOjRq5s0unrOyRJiw711m1oAMgj1oABIAgBDABBCGAACEIAA0AQAhgAghDAABCEAAaAIAQwAAQhgAEgCAEMAEEIYAAIQgADQBACGACCEMAAEIQABoAgBDAABCGAASAIAQwAQQhgAAhCAANAEAIYAIIQwAAQhAAGgCAEMAAEIYABIAgBDABBCGAACEIAA0AQAhgAghDAABCEAAaAIAQwAAQhgAEgCAEMAEEIYAAIQgADQBACGACCEMAAEIQABoAgBDAABCGAASBIXQJ4aGhIujBSVd9t27Zp27ZtNa4IAOI11mOQ06dPS+5V9S0WizWuBgDygSUIAAhCAANAEAIYAIIQwAAQhAAGgCAEMAAEIYABIAgBDABBCGAACEIAA0AQAhgAghDAABCEAAaAIAQwAAQhgAEgCAEMAEEIYAAIQgADQBACGACCEMAAEIQABoAgBDAABCGAASAIAQwAQQhgAAhCAANAEAIYAIIQwAAQhAAGgCAEMAAEIYABIAgBDABBCGAACEIAA0AQAhgAghDAABCEAAaAIAQwAATJXQAPDQ1p//79WrlyZVWXu+66S8ViUaVSSevXr9fKlSv17LPPqlQqqbu7W319fVqzZo36+/vV3d2tYrGo7u5ulUolFYtFtbe36+6779aGDRtUKpUq1lUqlXT//ffrgQceGHeM6RitaXS/iduzrZrjV+pT69pmImv+Ru+TPNWJ+aWWj4XcBfCxY8em1f/cuXPq6elRoVDQ0NCQJOmxxx5ToVDQwMCAHnnkEZ08eVJbt27VwMCAenp6NDAwoO3bt6unp0dnzpzR2bNndfjwYW3fvr3iOIVCQQcPHtSBAwfGHWM6Rmsa3W/i9myr5viV+tS6tpnImr/R+yRPdWJ+qeVjIVcB/NRTT81ov8HBQe3cuXNs2921a9cuubtOnDghSTpx4oTcXYODg3J39fb2anBwcNxxent7M5/lSqWSdu/ePW48d9eePXuqflYslUras2fP2H7FYnHc9mw/u04cr9LtyupTzb71ljV/5ffJ7t27c1En5pdaPxbCAviyMz9SsVjUpk2bxi47duyYteNfuHBhyuvPnz+f2Zb1LFcoFDQ8PDypfWRkpOpnxUKhMFbTyMiIenp6xm3P9rPrxPEq3a6sPtXsW29Z81d+n1S674BLUevHwkUD2Mw2mFmfmfUdPXp0VgfPo717905q27dvn9x9Uvvw8HBm/yz79u0bC4zh4WENDg6O2672ONWaOF6l25XVp5p96y1r/srvE3fPRZ2YX2r9WLhoALv759y9zd3bli1bNmsDX7jiJ9Xa2qrHH3987JIHd95556S2VatWycwmtTc2Nmb2z7Jq1So1NjaO7dfS0jJuu9rjVGvieJVuV1afavatt6z5K79PzCwXdWJ+qfVjIVdrwOvXr5+1Y1122dQ3bcGCBZltnZ2dk9q7urrG7oRyDQ0Nmf2zdHV1jdXU0NCgLVu2jNuu9jjVmjhepduV1aeafesta/7K75NK9x1wKWr9WMhVAN93330z2q+lpUXr1q0b2zYzrV27VmamxYsXS5IWL14sMxt75dTR0aGWlpZxx+no6FBTU9Ok4zc1NWn16tXjxjMztbe3Z/bP0tTUpPb29rH9Wltbx21Xe5xqTRyv0u3K6lPNvvWWNX/l98nq1atzUSfml1o/Fia/rAu2dOnSaX0UbeHChdqyZYuWLFmi/v5+DQ0NafPmzbrttts0ODiozs5Obd26VY8++qgKhYK6u7v1xBNPqLOzU8ePH9eDDz4od9eKFSumfHbr6urSkSNHZGbavHnz2DGmo6ura6ymrO3ZVs3xK/WpdW0zkTV/o/dJnurE/FLLx4JlvblUSVtbm/f19U17kDVr1ujEyVMa+YnlOn19hyRp0aFe3fLB5ZPWfjdt2iRJuVkTBoBLZWb97t42sT1XSxAA8F5CAANAEAIYAIIQwAAQhAAGgCAEMAAEIYABIAgBDABBCGAACEIAA0AQAhgAghDAABCEAAaAIAQwAAQhgAEgCAEMAEEIYAAIQgADQBACGACCEMAAEIQABoAgBDAABCGAASAIAQwAQQhgAAhCAANAEAIYAIIQwAAQhAAGgCAEMAAEIYABIAgBDABBCGAACEIAA0AQAhgAghDAABCEAAaAIAQwAARprMcgixYt0olTp6vq29raWuNqACAf6hLAzc3NOnr83ar6bty4scbVAEA+sAQBAEEIYAAIQgADQBACGACCEMAAEIQABoAgBDAABCGAASAIAQwAQQhgAAhCAANAEAIYAIIQwAAQhAAGgCAEMAAEIYABIAgBDABBCGAACEIAA0AQAhgAghDAABCEAAaAIAQwAAQhgAEgCAEMAEEIYAAIQgADQBACGACCEMAAEIQABoAgBDAABCGAASAIAQwAQQhgAAhCAANAEAIYAIIQwAAQhAAGgCAEMAAEIYABIEhj3UYaGVbDqZIWHeqVJDWcekfS8roNDwB5U5cAbm1t1dDQkCSpuXk0dJertbW1HsMDQC6Zu1fdua2tzfv6+mpYDgDMP2bW7+5tE9tZAwaAIAQwAAQhgAEgCAEMAEEIYAAIQgADQBACGACCEMAAEIQABoAgBDAABCGAASAIAQwAQQhgAAhCAANAEAIYAIIQwAAQhAAGgCAEMAAEIYABIAgBDABBpvVPOc3sqKTvz3CspZKOzXDfSNRdf3O1duqur7lU9wp3XzaxcVoBfCnMrC/rv4LmHXXX31ytnbrra67WXY4lCAAIQgADQJB6BvDn6jjWbKLu+purtVN3fc3VusfUbQ0YADAeSxAAEIQABoAgNQ9gM2s3s1fNrGhmn6r1eJfCzD5gZt80swNm9oqZbUrb32dme83sSPp1SXStWcyswcy+Y2a70u1rzeyFdO6/YmYLo2ucyMyuNrNnzOyQmR00s1+cC/NtZn+Q/oy8bGZfNrMr8jrfZvZFM3vbzF4ua8ucY0s8kd6Gl8zs5pzV/Zn0Z+UlM/u6mV1ddt1Dad2vmtndMVVPT00D2MwaJP21pNWSbpD0m2Z2Qy3HvETDkv7Q3W+QdKuk30vr/ZSk59z9OknPpdt5tEnSwbLtT0v6S3dvlXRc0idDqpra45L2uPv1km5SUn+u59vMmiV1S2pz9xslNUi6V/md7y9Jap/QVmmOV0u6Lr1skPRknWrM8iVNrnuvpBvd/cOSDkt6SJLSx+m9kn4u3edv0vzJtVq/Av4FSUV3/293PyfpaUn31HjMGXP3N939xfT7/1MSBs1Kai6k3QqSfjWmwsrM7BpJayR9Pt02SR+X9EzaJXd1m9lPSfplSV+QJHc/5+4/1ByYb0mNkhaZWaOkKyW9qZzOt7v/u6R3JjRXmuN7JG33xLckXW1mP1OfSsfLqtvd/9Xdh9PNb0m6Jv3+HklPu/tZd/+epKKS/Mm1Wgdws6TXy7bfSNtyz8xaJH1U0guSlrv7m+lVb0laHlTWVP5K0h9LupBuN0n6YdkPax7n/lpJRyX9bbp08nkzu0o5n293H5L055L+R0nwviupX/mf73KV5nguPWZ/R9Lu9Pu5VPcY3oTLYGaLJf2DpN939x+VX+fJ5/Zy9dk9M1sr6W1374+uZZoaJd0s6Ul3/6ikk5qw3JDT+V6i5BXXtZLeL+kqTf5Vec7I4xxfjJk9rGTJcEd0LZei1gE8JOkDZdvXpG25ZWYLlITvDnf/Wtr8g9Ffw9Kvb0fVV8EvSVpnZoNKlnk+rmRt9er0V2Qpn3P/hqQ33P2FdPsZJYGc9/leJel77n7U3c9L+pqS+yDv812u0hzn/jFrZr8taa2k9f7jP2TIfd1Zah3A/yXpuvTd4YVKFsl31njMGUvXTb8g6aC7P1Z21U5JXen3XZL+qd61TcXdH3L3a9y9Rckcf8Pd10v6pqRfS7vlse63JL1uZh9Km35F0gHlfL6VLD3camZXpj8zo3Xner4nqDTHOyV1pp+GuFXSu2VLFeHMrF3JUts6dz9VdtVOSfea2eVmdq2SNxG/HVHjtLh7TS+SOpS8W/mapIdrPd4l1nq7kl/FXpL03fTSoWQ99TlJRyTtk/S+6FqnuA0rJe1Kv/+gkh/CoqSvSro8ur6Mej8iqS+d83+UtGQuzLekRyUdkvSypL+TdHle51vSl5WsVZ9X8lvHJyvNsSRT8sml1yQNKPmkR57qLipZ6x19fH62rP/Dad2vSlodPe/VXPhTZAAIwptwABCEAAaAIAQwAAQhgAEgCAEMAEEIYIQxs582s6fN7DUz6zezXjP72WkeY9DMlla4boGZvTjVWGbWUn62LaCeGi/eBZh96R8wfF1Swd3vTdtuUnJOgsOzNMztkv7zImO9PsX+QE3xChhR7pB03t0/O9rg7vslfcfMnjOzF81swMzukSQzu8rM/tnM9qfn4P2NsmNtLOt/fVl7u5KTtWSO5e7/UdNbCFwEAYwoNyo5g9hEZyR9wt1vVhKcf5G+gm2X9L/ufpMn5+DdU7bPsbT/k5L+qKz9Dkn/NsVYQCgCGHljkv7UzF5S8ieyzUqWCgYk3Wlmnzazj7n7u2X7jJ40qV9SizR20vR3fPz5AoBcIYAR5RVJt2S0r5e0TNIt7v4RST+QdIW7H1ZyprQBST1m9idl+5xNv47ox+9rtEv6l4uMBYQigBHlG5IuN7MNow1m9mFJK5Sc2/i8md2RbsvM3i/plLv/vaTPKAnjqYyu/1Ycy8w+Nmu3BpgBAhghPDkL1CckrUo/GvaKpD+T1CupzcwGJHUqOeOYJP28pG+b2XclbZXUU+nY6f8Ca3X3QxcZ6610lw+Z2Rtll1+f9RsMZOBsaJh3zOx2Sb/l7r8bXQswFQIYAIKwBAEAQQhgAAhCAANAEAIYAIIQwAAQhAAGgCD/D6qn7F8JhBDJAAAAAElFTkSuQmCC\n","text/plain":["<Figure size 432x288 with 1 Axes>"]},"metadata":{"tags":[],"needs_background":"light"}}]},{"cell_type":"code","metadata":{"colab":{"base_uri":"https://localhost:8080/"},"id":"Hf6X7NfBfsYm","executionInfo":{"status":"ok","timestamp":1625713109315,"user_tz":240,"elapsed":58,"user":{"displayName":"Tzu-Hua Huang","photoUrl":"https://lh3.googleusercontent.com/a-/AOh14GgbXOdZWwXGVTDuMHbeSme5C5oKAeF6I6ojY6qc=s64","userId":"01458545184256376706"}},"outputId":"f3dc5653-c176-47e4-c29a-e7c8b1d57a79"},"source":["# deal with outliers by winzoring\n","print(df1['Cash/CL'].quantile(0.00))\n","print(df1['Cash/CL'].quantile(0.90))"],"execution_count":189,"outputs":[{"output_type":"stream","text":["0.0\n","2.7112986730476494\n"],"name":"stdout"}]},{"cell_type":"code","metadata":{"colab":{"base_uri":"https://localhost:8080/"},"id":"rRZq1Po1y-R2","executionInfo":{"status":"ok","timestamp":1625713109317,"user_tz":240,"elapsed":51,"user":{"displayName":"Tzu-Hua Huang","photoUrl":"https://lh3.googleusercontent.com/a-/AOh14GgbXOdZWwXGVTDuMHbeSme5C5oKAeF6I6ojY6qc=s64","userId":"01458545184256376706"}},"outputId":"15cccfa6-cd49-4615-b806-023307b50f26"},"source":["# remove the effect of outliers\n","# according to the boxplot, replace 5% of the lowest and 2% the highest values by winsorizing\n","df1['Cash/CL'].clip(lower=df1['Cash/CL'].quantile(0.00),\n","                    upper=df1['Cash/CL'].quantile(0.90),\n","                    axis=0, inplace=True)\n","df1['Cash/CL'].describe()"],"execution_count":190,"outputs":[{"output_type":"execute_result","data":{"text/plain":["count    2022.000000\n","mean        1.028216\n","std         0.930716\n","min         0.000000\n","25%         0.221913\n","50%         0.642435\n","75%         2.083432\n","max         2.711299\n","Name: Cash/CL, dtype: float64"]},"metadata":{"tags":[]},"execution_count":190}]},{"cell_type":"code","metadata":{"colab":{"base_uri":"https://localhost:8080/","height":333},"id":"AkO4GSsDfsYn","executionInfo":{"status":"ok","timestamp":1625713109318,"user_tz":240,"elapsed":46,"user":{"displayName":"Tzu-Hua Huang","photoUrl":"https://lh3.googleusercontent.com/a-/AOh14GgbXOdZWwXGVTDuMHbeSme5C5oKAeF6I6ojY6qc=s64","userId":"01458545184256376706"}},"outputId":"b12afac3-6230-4532-829a-76234133e493"},"source":["# check if outliers exist after winzoring\n","sns.boxplot(df1['Cash/CL'])\n","plt.show()"],"execution_count":191,"outputs":[{"output_type":"stream","text":["/usr/local/lib/python3.7/dist-packages/seaborn/_decorators.py:43: FutureWarning: Pass the following variable as a keyword arg: x. From version 0.12, the only valid positional argument will be `data`, and passing other arguments without an explicit keyword will result in an error or misinterpretation.\n","  FutureWarning\n"],"name":"stderr"},{"output_type":"display_data","data":{"image/png":"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\n","text/plain":["<Figure size 432x288 with 1 Axes>"]},"metadata":{"tags":[],"needs_background":"light"}}]},{"cell_type":"markdown","metadata":{"id":"cDVxUzACfsYn"},"source":["##### Cash/Interest Expense"]},{"cell_type":"code","metadata":{"colab":{"base_uri":"https://localhost:8080/","height":333},"id":"IWd_Ffk5fsYn","executionInfo":{"status":"ok","timestamp":1625713109318,"user_tz":240,"elapsed":38,"user":{"displayName":"Tzu-Hua Huang","photoUrl":"https://lh3.googleusercontent.com/a-/AOh14GgbXOdZWwXGVTDuMHbeSme5C5oKAeF6I6ojY6qc=s64","userId":"01458545184256376706"}},"outputId":"1f40ce3d-e1c6-4364-a61e-94e009dd990f"},"source":["# check outliers with boxplot\n","sns.boxplot(df1['Cash/Interest Expense'])\n","plt.show()"],"execution_count":192,"outputs":[{"output_type":"stream","text":["/usr/local/lib/python3.7/dist-packages/seaborn/_decorators.py:43: FutureWarning: Pass the following variable as a keyword arg: x. From version 0.12, the only valid positional argument will be `data`, and passing other arguments without an explicit keyword will result in an error or misinterpretation.\n","  FutureWarning\n"],"name":"stderr"},{"output_type":"display_data","data":{"image/png":"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\n","text/plain":["<Figure size 432x288 with 1 Axes>"]},"metadata":{"tags":[],"needs_background":"light"}}]},{"cell_type":"code","metadata":{"colab":{"base_uri":"https://localhost:8080/"},"id":"5J6populfsYn","executionInfo":{"status":"ok","timestamp":1625713109319,"user_tz":240,"elapsed":32,"user":{"displayName":"Tzu-Hua Huang","photoUrl":"https://lh3.googleusercontent.com/a-/AOh14GgbXOdZWwXGVTDuMHbeSme5C5oKAeF6I6ojY6qc=s64","userId":"01458545184256376706"}},"outputId":"42d734ae-74d4-4c8c-dff9-136556ad9f55"},"source":["# deal with outliers by winzoring\n","print(df1['Cash/Interest Expense'].quantile(0.00))\n","print(df1['Cash/Interest Expense'].quantile(0.85))"],"execution_count":193,"outputs":[{"output_type":"stream","text":["0.0\n","47.27893278685917\n"],"name":"stdout"}]},{"cell_type":"code","metadata":{"colab":{"base_uri":"https://localhost:8080/"},"id":"l4_5jGRAzW8X","executionInfo":{"status":"ok","timestamp":1625713109320,"user_tz":240,"elapsed":28,"user":{"displayName":"Tzu-Hua Huang","photoUrl":"https://lh3.googleusercontent.com/a-/AOh14GgbXOdZWwXGVTDuMHbeSme5C5oKAeF6I6ojY6qc=s64","userId":"01458545184256376706"}},"outputId":"11ee5d08-35e6-45ea-c8ab-381933008b86"},"source":["# remove the effect of outliers\n","# according to the boxplot, replace 5% of the lowest and 3% of the highest values by winsorizing\n","df1['Cash/Interest Expense'].clip(lower=df1['Cash/Interest Expense'].quantile(0.00),\n","                                  upper=df1['Cash/Interest Expense'].quantile(0.85),\n","                                  axis=0, inplace=True)\n","df1['Cash/Interest Expense'].describe()"],"execution_count":194,"outputs":[{"output_type":"execute_result","data":{"text/plain":["count    2022.000000\n","mean       14.376623\n","std        17.010703\n","min         0.000000\n","25%         1.519326\n","50%         5.809825\n","75%        22.681178\n","max        47.278933\n","Name: Cash/Interest Expense, dtype: float64"]},"metadata":{"tags":[]},"execution_count":194}]},{"cell_type":"code","metadata":{"colab":{"base_uri":"https://localhost:8080/","height":333},"id":"lUkgx5FNfsYn","executionInfo":{"status":"ok","timestamp":1625713109730,"user_tz":240,"elapsed":433,"user":{"displayName":"Tzu-Hua Huang","photoUrl":"https://lh3.googleusercontent.com/a-/AOh14GgbXOdZWwXGVTDuMHbeSme5C5oKAeF6I6ojY6qc=s64","userId":"01458545184256376706"}},"outputId":"1c9e3d62-c514-4fdb-fbb0-288b23b5b7c3"},"source":["# check if outliers exist after winzoring\n","sns.boxplot(df1['Cash/Interest Expense'])\n","plt.show()"],"execution_count":195,"outputs":[{"output_type":"stream","text":["/usr/local/lib/python3.7/dist-packages/seaborn/_decorators.py:43: FutureWarning: Pass the following variable as a keyword arg: x. From version 0.12, the only valid positional argument will be `data`, and passing other arguments without an explicit keyword will result in an error or misinterpretation.\n","  FutureWarning\n"],"name":"stderr"},{"output_type":"display_data","data":{"image/png":"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\n","text/plain":["<Figure size 432x288 with 1 Axes>"]},"metadata":{"tags":[],"needs_background":"light"}}]},{"cell_type":"markdown","metadata":{"id":"ImnjJR0-fsYn"},"source":["##### Quick Ratio"]},{"cell_type":"code","metadata":{"colab":{"base_uri":"https://localhost:8080/","height":333},"id":"ZYwqpQaQfsYo","executionInfo":{"status":"ok","timestamp":1625713109784,"user_tz":240,"elapsed":321,"user":{"displayName":"Tzu-Hua Huang","photoUrl":"https://lh3.googleusercontent.com/a-/AOh14GgbXOdZWwXGVTDuMHbeSme5C5oKAeF6I6ojY6qc=s64","userId":"01458545184256376706"}},"outputId":"de8449b7-4dc2-4bb8-e844-9a35c7fa460c"},"source":["# check outliers with boxplot\n","sns.boxplot(df1['Quick Ratio'])\n","plt.show()"],"execution_count":196,"outputs":[{"output_type":"stream","text":["/usr/local/lib/python3.7/dist-packages/seaborn/_decorators.py:43: FutureWarning: Pass the following variable as a keyword arg: x. From version 0.12, the only valid positional argument will be `data`, and passing other arguments without an explicit keyword will result in an error or misinterpretation.\n","  FutureWarning\n"],"name":"stderr"},{"output_type":"display_data","data":{"image/png":"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\n","text/plain":["<Figure size 432x288 with 1 Axes>"]},"metadata":{"tags":[],"needs_background":"light"}}]},{"cell_type":"code","metadata":{"colab":{"base_uri":"https://localhost:8080/"},"id":"VtQgyPIifsYo","executionInfo":{"status":"ok","timestamp":1625713109786,"user_tz":240,"elapsed":312,"user":{"displayName":"Tzu-Hua Huang","photoUrl":"https://lh3.googleusercontent.com/a-/AOh14GgbXOdZWwXGVTDuMHbeSme5C5oKAeF6I6ojY6qc=s64","userId":"01458545184256376706"}},"outputId":"32db7a2e-50e3-49c4-e3d7-930398e7ba1c"},"source":["# deal with outliers by winzoring\n","print(df1['Quick Ratio'].quantile(0.00))\n","print(df1['Quick Ratio'].quantile(0.95))"],"execution_count":197,"outputs":[{"output_type":"stream","text":["0.056552262815804\n","5.555035738890612\n"],"name":"stdout"}]},{"cell_type":"code","metadata":{"colab":{"base_uri":"https://localhost:8080/"},"id":"Hoh7l_AwzpY4","executionInfo":{"status":"ok","timestamp":1625713109787,"user_tz":240,"elapsed":303,"user":{"displayName":"Tzu-Hua Huang","photoUrl":"https://lh3.googleusercontent.com/a-/AOh14GgbXOdZWwXGVTDuMHbeSme5C5oKAeF6I6ojY6qc=s64","userId":"01458545184256376706"}},"outputId":"cb9dacb7-f3e1-4a94-96dd-37c3ceaf49aa"},"source":["# remove the effect of outliers\n","# according to the boxplot, replace 5% of the highest values by winsorizing\n","df1['Quick Ratio'].clip(lower=df1['Quick Ratio'].quantile(0.00),\n","                        upper=df1['Quick Ratio'].quantile(0.95),\n","                        axis=0, inplace=True)\n","df1['Quick Ratio'].describe()"],"execution_count":198,"outputs":[{"output_type":"execute_result","data":{"text/plain":["count    2022.000000\n","mean        1.949707\n","std         1.336237\n","min         0.056552\n","25%         0.941378\n","50%         1.549811\n","75%         2.855418\n","max         5.555036\n","Name: Quick Ratio, dtype: float64"]},"metadata":{"tags":[]},"execution_count":198}]},{"cell_type":"code","metadata":{"colab":{"base_uri":"https://localhost:8080/","height":333},"id":"BPbqkv1vfsYo","executionInfo":{"status":"ok","timestamp":1625713109787,"user_tz":240,"elapsed":292,"user":{"displayName":"Tzu-Hua Huang","photoUrl":"https://lh3.googleusercontent.com/a-/AOh14GgbXOdZWwXGVTDuMHbeSme5C5oKAeF6I6ojY6qc=s64","userId":"01458545184256376706"}},"outputId":"f91c8a48-247b-4c1c-c14d-33cb92ccc721"},"source":["# check if outliers exist after winzoring\n","sns.boxplot(df1['Quick Ratio'])\n","plt.show()"],"execution_count":199,"outputs":[{"output_type":"stream","text":["/usr/local/lib/python3.7/dist-packages/seaborn/_decorators.py:43: FutureWarning: Pass the following variable as a keyword arg: x. From version 0.12, the only valid positional argument will be `data`, and passing other arguments without an explicit keyword will result in an error or misinterpretation.\n","  FutureWarning\n"],"name":"stderr"},{"output_type":"display_data","data":{"image/png":"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\n","text/plain":["<Figure size 432x288 with 1 Axes>"]},"metadata":{"tags":[],"needs_background":"light"}}]},{"cell_type":"markdown","metadata":{"id":"qDXIcwOGfsYq"},"source":["##### Long-term Debt/TA"]},{"cell_type":"code","metadata":{"colab":{"base_uri":"https://localhost:8080/","height":333},"id":"fqNYvjaIfsYq","executionInfo":{"status":"ok","timestamp":1625713109790,"user_tz":240,"elapsed":282,"user":{"displayName":"Tzu-Hua Huang","photoUrl":"https://lh3.googleusercontent.com/a-/AOh14GgbXOdZWwXGVTDuMHbeSme5C5oKAeF6I6ojY6qc=s64","userId":"01458545184256376706"}},"outputId":"d845453f-9839-4680-b350-f6ca7ad05cd5"},"source":["# check outliers with boxplot\n","sns.boxplot(df1['Long-term Debt/TA'])\n","plt.show()"],"execution_count":200,"outputs":[{"output_type":"stream","text":["/usr/local/lib/python3.7/dist-packages/seaborn/_decorators.py:43: FutureWarning: Pass the following variable as a keyword arg: x. From version 0.12, the only valid positional argument will be `data`, and passing other arguments without an explicit keyword will result in an error or misinterpretation.\n","  FutureWarning\n"],"name":"stderr"},{"output_type":"display_data","data":{"image/png":"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\n","text/plain":["<Figure size 432x288 with 1 Axes>"]},"metadata":{"tags":[],"needs_background":"light"}}]},{"cell_type":"code","metadata":{"colab":{"base_uri":"https://localhost:8080/"},"id":"sYpH1RssfsYq","executionInfo":{"status":"ok","timestamp":1625713109791,"user_tz":240,"elapsed":277,"user":{"displayName":"Tzu-Hua Huang","photoUrl":"https://lh3.googleusercontent.com/a-/AOh14GgbXOdZWwXGVTDuMHbeSme5C5oKAeF6I6ojY6qc=s64","userId":"01458545184256376706"}},"outputId":"ab37d681-2391-4761-fb92-eaef9726916d"},"source":["# deal with outliers by winzoring\n","print(df1['Long-term Debt/TA'].quantile(0.00))\n","print(df1['Long-term Debt/TA'].quantile(0.97))"],"execution_count":201,"outputs":[{"output_type":"stream","text":["0.0\n","0.7796233351806668\n"],"name":"stdout"}]},{"cell_type":"code","metadata":{"colab":{"base_uri":"https://localhost:8080/"},"id":"LpnxHX22z7te","executionInfo":{"status":"ok","timestamp":1625713109792,"user_tz":240,"elapsed":270,"user":{"displayName":"Tzu-Hua Huang","photoUrl":"https://lh3.googleusercontent.com/a-/AOh14GgbXOdZWwXGVTDuMHbeSme5C5oKAeF6I6ojY6qc=s64","userId":"01458545184256376706"}},"outputId":"c3e4a657-7129-4ca8-d7ad-f4a676740a07"},"source":["# remove the effect of outliers\n","# according to the boxplot, replace 5% of the highest values by winsorizing\n","df1['Long-term Debt/TA'].clip(lower=df1['Long-term Debt/TA'].quantile(0.00),\n","                              upper=df1['Long-term Debt/TA'].quantile(0.97),\n","                              axis=0, inplace=True)\n","df1['Long-term Debt/TA'].describe()"],"execution_count":202,"outputs":[{"output_type":"execute_result","data":{"text/plain":["count    2022.000000\n","mean        0.312041\n","std         0.201288\n","min         0.000000\n","25%         0.153549\n","50%         0.303907\n","75%         0.435809\n","max         0.779623\n","Name: Long-term Debt/TA, dtype: float64"]},"metadata":{"tags":[]},"execution_count":202}]},{"cell_type":"code","metadata":{"colab":{"base_uri":"https://localhost:8080/","height":333},"id":"K9XM_5XPfsYq","executionInfo":{"status":"ok","timestamp":1625713109792,"user_tz":240,"elapsed":261,"user":{"displayName":"Tzu-Hua Huang","photoUrl":"https://lh3.googleusercontent.com/a-/AOh14GgbXOdZWwXGVTDuMHbeSme5C5oKAeF6I6ojY6qc=s64","userId":"01458545184256376706"}},"outputId":"b2dc499a-0cc8-451d-ed7b-2e3972acf60c"},"source":["# check if outliers exist after winzoring\n","sns.boxplot(df1['Long-term Debt/TA'])\n","plt.show()"],"execution_count":203,"outputs":[{"output_type":"stream","text":["/usr/local/lib/python3.7/dist-packages/seaborn/_decorators.py:43: FutureWarning: Pass the following variable as a keyword arg: x. From version 0.12, the only valid positional argument will be `data`, and passing other arguments without an explicit keyword will result in an error or misinterpretation.\n","  FutureWarning\n"],"name":"stderr"},{"output_type":"display_data","data":{"image/png":"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\n","text/plain":["<Figure size 432x288 with 1 Axes>"]},"metadata":{"tags":[],"needs_background":"light"}}]},{"cell_type":"markdown","metadata":{"id":"ozRxzvD3fsYr"},"source":["##### Fixed Assets/TA"]},{"cell_type":"code","metadata":{"colab":{"base_uri":"https://localhost:8080/","height":333},"id":"SOn-yub2fsYr","executionInfo":{"status":"ok","timestamp":1625713110087,"user_tz":240,"elapsed":547,"user":{"displayName":"Tzu-Hua Huang","photoUrl":"https://lh3.googleusercontent.com/a-/AOh14GgbXOdZWwXGVTDuMHbeSme5C5oKAeF6I6ojY6qc=s64","userId":"01458545184256376706"}},"outputId":"5ae8cb23-1ee2-4912-a764-1986e0852d6c"},"source":["# check outliers with boxplot\n","sns.boxplot(df1['Fixed Assets/TA'])\n","plt.show()"],"execution_count":204,"outputs":[{"output_type":"stream","text":["/usr/local/lib/python3.7/dist-packages/seaborn/_decorators.py:43: FutureWarning: Pass the following variable as a keyword arg: x. From version 0.12, the only valid positional argument will be `data`, and passing other arguments without an explicit keyword will result in an error or misinterpretation.\n","  FutureWarning\n"],"name":"stderr"},{"output_type":"display_data","data":{"image/png":"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\n","text/plain":["<Figure size 432x288 with 1 Axes>"]},"metadata":{"tags":[],"needs_background":"light"}}]},{"cell_type":"code","metadata":{"colab":{"base_uri":"https://localhost:8080/"},"id":"Ifzf2Xm9fsYr","executionInfo":{"status":"ok","timestamp":1625713110088,"user_tz":240,"elapsed":41,"user":{"displayName":"Tzu-Hua Huang","photoUrl":"https://lh3.googleusercontent.com/a-/AOh14GgbXOdZWwXGVTDuMHbeSme5C5oKAeF6I6ojY6qc=s64","userId":"01458545184256376706"}},"outputId":"6fa320d5-e260-4ccd-89b0-b40e234ab878"},"source":["# deal with outliers by winzoring\n","print(df1['Fixed Assets/TA'].quantile(0.00))\n","print(df1['Fixed Assets/TA'].quantile(0.95))"],"execution_count":205,"outputs":[{"output_type":"stream","text":["0.0\n","0.7986178456391186\n"],"name":"stdout"}]},{"cell_type":"code","metadata":{"colab":{"base_uri":"https://localhost:8080/"},"id":"pNLOqRDq0I7s","executionInfo":{"status":"ok","timestamp":1625713110088,"user_tz":240,"elapsed":28,"user":{"displayName":"Tzu-Hua Huang","photoUrl":"https://lh3.googleusercontent.com/a-/AOh14GgbXOdZWwXGVTDuMHbeSme5C5oKAeF6I6ojY6qc=s64","userId":"01458545184256376706"}},"outputId":"69396d2e-f5d1-4698-b80b-808b92149be7"},"source":["# remove the effect of outliers\n","# according to the boxplot, replace 5% of the highest values by winsorizing\n","df1['Fixed Assets/TA'].clip(lower=df1['Fixed Assets/TA'].quantile(0.00),\n","                            upper=df1['Fixed Assets/TA'].quantile(0.95),\n","                            axis=0, inplace=True)\n","df1['Fixed Assets/TA'].describe()"],"execution_count":206,"outputs":[{"output_type":"execute_result","data":{"text/plain":["count    2022.000000\n","mean        0.250316\n","std         0.231879\n","min         0.000000\n","25%         0.077808\n","50%         0.174078\n","75%         0.353765\n","max         0.798618\n","Name: Fixed Assets/TA, dtype: float64"]},"metadata":{"tags":[]},"execution_count":206}]},{"cell_type":"code","metadata":{"colab":{"base_uri":"https://localhost:8080/","height":333},"id":"PS0TVCtrfsYs","executionInfo":{"status":"ok","timestamp":1625713110338,"user_tz":240,"elapsed":269,"user":{"displayName":"Tzu-Hua Huang","photoUrl":"https://lh3.googleusercontent.com/a-/AOh14GgbXOdZWwXGVTDuMHbeSme5C5oKAeF6I6ojY6qc=s64","userId":"01458545184256376706"}},"outputId":"91429ccf-314c-46f1-fe44-18c239253eb1"},"source":["# check if outliers exist after winzoring\n","sns.boxplot(df1['Fixed Assets/TA'])\n","plt.show()"],"execution_count":207,"outputs":[{"output_type":"stream","text":["/usr/local/lib/python3.7/dist-packages/seaborn/_decorators.py:43: FutureWarning: Pass the following variable as a keyword arg: x. From version 0.12, the only valid positional argument will be `data`, and passing other arguments without an explicit keyword will result in an error or misinterpretation.\n","  FutureWarning\n"],"name":"stderr"},{"output_type":"display_data","data":{"image/png":"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\n","text/plain":["<Figure size 432x288 with 1 Axes>"]},"metadata":{"tags":[],"needs_background":"light"}}]},{"cell_type":"markdown","metadata":{"id":"xio6LJ9A0g6z"},"source":["##### Sales per Employee"]},{"cell_type":"code","metadata":{"colab":{"base_uri":"https://localhost:8080/","height":333},"id":"uSMQvWCr0g6z","executionInfo":{"status":"ok","timestamp":1625713110339,"user_tz":240,"elapsed":67,"user":{"displayName":"Tzu-Hua Huang","photoUrl":"https://lh3.googleusercontent.com/a-/AOh14GgbXOdZWwXGVTDuMHbeSme5C5oKAeF6I6ojY6qc=s64","userId":"01458545184256376706"}},"outputId":"7cdf9d5d-8e8f-475c-9be5-f96f06c9736f"},"source":["# check outliers with boxplot\n","sns.boxplot(df1['Sales per Employee'])\n","plt.show()"],"execution_count":208,"outputs":[{"output_type":"stream","text":["/usr/local/lib/python3.7/dist-packages/seaborn/_decorators.py:43: FutureWarning: Pass the following variable as a keyword arg: x. From version 0.12, the only valid positional argument will be `data`, and passing other arguments without an explicit keyword will result in an error or misinterpretation.\n","  FutureWarning\n"],"name":"stderr"},{"output_type":"display_data","data":{"image/png":"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\n","text/plain":["<Figure size 432x288 with 1 Axes>"]},"metadata":{"tags":[],"needs_background":"light"}}]},{"cell_type":"code","metadata":{"colab":{"base_uri":"https://localhost:8080/"},"id":"fQydIKg-0g60","executionInfo":{"status":"ok","timestamp":1625713110339,"user_tz":240,"elapsed":50,"user":{"displayName":"Tzu-Hua Huang","photoUrl":"https://lh3.googleusercontent.com/a-/AOh14GgbXOdZWwXGVTDuMHbeSme5C5oKAeF6I6ojY6qc=s64","userId":"01458545184256376706"}},"outputId":"894f5212-dfef-4005-c18e-1b2f6d5c32fb"},"source":["# deal with outliers by winzoring\n","print(df1['Sales per Employee'].quantile(0.00))\n","print(df1['Sales per Employee'].quantile(0.90))"],"execution_count":209,"outputs":[{"output_type":"stream","text":["0.0\n","2.0139697494383135\n"],"name":"stdout"}]},{"cell_type":"code","metadata":{"colab":{"base_uri":"https://localhost:8080/"},"id":"kAZCO9ge0g60","executionInfo":{"status":"ok","timestamp":1625713110342,"user_tz":240,"elapsed":43,"user":{"displayName":"Tzu-Hua Huang","photoUrl":"https://lh3.googleusercontent.com/a-/AOh14GgbXOdZWwXGVTDuMHbeSme5C5oKAeF6I6ojY6qc=s64","userId":"01458545184256376706"}},"outputId":"b980fe8e-f8ff-4b08-d6d9-52f5c09259ea"},"source":["# remove the effect of outliers\n","# according to the boxplot, replace 5% of the highest values by winsorizing\n","df1['Sales per Employee'].clip(lower=df1['Sales per Employee'].quantile(0.00),\n","                               upper=df1['Sales per Employee'].quantile(0.90),\n","                               axis=0, inplace=True)\n","df1['Sales per Employee'].describe()"],"execution_count":210,"outputs":[{"output_type":"execute_result","data":{"text/plain":["count    2022.000000\n","mean        0.624221\n","std         0.594010\n","min         0.000000\n","25%         0.224528\n","50%         0.379812\n","75%         0.801262\n","max         2.013970\n","Name: Sales per Employee, dtype: float64"]},"metadata":{"tags":[]},"execution_count":210}]},{"cell_type":"code","metadata":{"colab":{"base_uri":"https://localhost:8080/","height":333},"id":"s7M1FnIs0g60","executionInfo":{"status":"ok","timestamp":1625713110351,"user_tz":240,"elapsed":45,"user":{"displayName":"Tzu-Hua Huang","photoUrl":"https://lh3.googleusercontent.com/a-/AOh14GgbXOdZWwXGVTDuMHbeSme5C5oKAeF6I6ojY6qc=s64","userId":"01458545184256376706"}},"outputId":"1ecb47cf-44fc-4d53-ab82-7cce57f87f0b"},"source":["# check if outliers exist after winzoring\n","sns.boxplot(df1['Sales per Employee'])\n","plt.show()"],"execution_count":211,"outputs":[{"output_type":"stream","text":["/usr/local/lib/python3.7/dist-packages/seaborn/_decorators.py:43: FutureWarning: Pass the following variable as a keyword arg: x. From version 0.12, the only valid positional argument will be `data`, and passing other arguments without an explicit keyword will result in an error or misinterpretation.\n","  FutureWarning\n"],"name":"stderr"},{"output_type":"display_data","data":{"image/png":"iVBORw0KGgoAAAANSUhEUgAAAWAAAAEGCAYAAABbzE8LAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADh0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uMy4yLjIsIGh0dHA6Ly9tYXRwbG90bGliLm9yZy+WH4yJAAASL0lEQVR4nO3df5BV5X3H8c8XlsjiJgpird0oV1yNxWD8QW21phVbJwtWTRszMWMqKzqYkgA1Yxoz3Eamk047zdREsSa1GaOOjtpoTLWDRBrBtrH+WBwBbdAcdWlKTdRVMQii4Ld/nGev5y532btw7/0u8H7N7Oy95zznnO8+99zPnn3uvc+auwsA0HpjogsAgP0VAQwAQQhgAAhCAANAEAIYAIK0jaTx5MmTvVQqNakUANg3rV69+lV3P3Tw8hEFcKlUUm9vb+OqAoD9gJltqLWcIQgACEIAA0AQAhgAghDAABCEAAaAIAQwAAQhgAEgCAEMAEEIYAAIQgADQBACGACCEMAAEIQABoAgBDAABCGAASAIAQwAQQhgAAhCAANAEAIYAIKM6H/C7e2WLl2qLMuasu+NGzdKkjo7O5uy/71BV1eXFixYEF0GsNfYrwI4yzI99fRPtWPCpIbve+yWTZKkX2zbr7q0YuyW16JLAPY6+11a7JgwSVuPm93w/bavXyZJTdn33mDg5wdQP8aAASAIAQwAQQhgAAhCAANAEAIYAIIQwAAQhAAGgCAEMAAEIYABIAgBDABBCGAACEIAA0AQAhgAghDAABCEAAaAIAQwAAQhgAEgCAEMAEEIYAAIQgADQBACGACCEMAAEIQABoAgBDAABCGAASAIAQwAQQhgAAhCAANAEAIYAIIQwAAQhAAGgCAEMAAEIYABIAgBDABBCGAACEIAA0AQAhgAghDAABCEAAaAIAQwAAQhgAEgSEsCeOnSpVq6dGkrDgUgEM/1kWlrxUGyLGvFYQAE47k+MgxBAEAQAhgAghDAABCEAAaAIAQwAAQhgAEgCAEMAEEIYAAIQgADQBACGACCEMAAEIQABoAgBDAABCGAASAIAQwAQQhgAAhCAANAEAIYAIIQwAAQhAAGgCAEMAAEIYABIAgBDABBCGAACEIAA0AQAhgAghDAABCEAAaAIAQwAAQhgAEgCAEMAEEIYAAIQgADQBACGACCEMAAEIQABoAgBDAABCGAASAIAQyg4bIs06xZszRv3jxlWabLLrtMs2fPVm9vr+bOnauZM2fqjjvu0FlnnaWVK1dq4cKFyrKs8r2np0czZ87URRddpFmzZinLMmVZpu7ubnV3dyvLMvX391dt19/fP2Qdc+fO1fz586vaZlmmc845R1mWSZL6+/s1f/589fT0VGpduHChzjzzzMpXo5m71914xowZ3tvbO+KDLFq0SJJ07bXXjnjbRlq0aJFWv/BLbT1udsP33b5+mSQ1Zd97g/b1y3TK1MPCH2PEGniub9q0SX19fZKkUqlUud3R0aHNmzdXbdPW1qYdO3ZoypQp2rBhg6ZMmVJpP6BUKklS1T5POOEE3X///ZXtzjvvPF1xxRVV2/X09FTtq1QqVdquWbNGfX19KpVKuvnmm3XNNdfovvvuq7Tt6OjQW2+9pWJGrlq1arf6xcxWu/uMwcu5AgbQUFu2bKkKveLtweErSdu3b5e7q6+vr/J9sL6+vp32+cADD1Rtt3z58qqr4CzLdtrXQNtly5ZV1vX19am3t1fLly+vart582YNvkBt9FVwW0P3NoSNGzdq69atld+OUbIs05h36r/iR/3GvP2msuxX4Y8xYmVZpq1bt7bkWO+++27V/R07dujWW2+tXAV//etfr3vbJUuW7LSsFYa9AjazeWbWa2a9r7zySitqArAXe++990KOu337dq1YsaJyv9aV9FBqXe22wrBXwO5+o6QbpXwMeHcO0tnZKWn0jAGj8d4b/yF1MQa831u0aJHWr1+vbdu2tfzYbW1tOvvssyv3i2PPw6k13tsKjAEDaKgjjzyyJccZN25c1f2xY8fq4osvrtwvl8t1b7tkyZKdlrUCAQygoSZMmFB514KkqtsdHR07tW9ra5OZqVQqVb4PViqVdtrnrFmzqrbr7u7WIYccUmnT1dW1074G2s6ePbuyrlQqacaMGeru7q5q29HRITOrWra774IYCgEMoOHK5bLa29t17LHHqlwuq6urSxMmTNCSJUs0depUmZkuv/xyjRkzRosXL9b06dNVLpcr3weCsrOzU+3t7SqXyyqXyxo/frzGjx+vcrmsOXPmVG1XvPodXMfUqVM1bdq0qrblclkHHnhg5Up5zpw5mjZtmkqlUqXW6dOnN7WfeB9wg/A+YN4HjNHzXB9teB8wAIwyBDAABCGAASAIAQwAQQhgAAhCAANAEAIYAIIQwAAQhAAGgCAEMAAEIYABIAgBDABBCGAACEIAA0AQAhgAghDAABCEAAaAIAQwAAQhgAEgCAEMAEEIYAAIQgADQBACGACCEMAAEIQABoAgBDAABCGAASAIAQwAQQhgAAhCAANAEAIYAIIQwAAQhAAGgCAEMAAEIYABIAgBDABBCGAACEIAA0CQtlYcpKurqxWHARCM5/rItCSAFyxY0IrDAAjGc31kGIIAgCAEMAAEIYABIAgBDABBCGAACEIAA0AQAhgAghDAABCEAAaAIAQwAAQhgAEgCAEMAEEIYAAIQgADQBACGACCEMAAEIQABoAgBDAABCGAASAIAQwAQQhgAAhCAANAEAIYAIIQwAAQhAAGgCAEMAAEIYABIAgBDABBCGAACEIAA0AQAhgAghDAABCEAAaAIAQwAAQhgAEgCAEMAEEIYAAIQgADQBACGACCEMAAEKQtuoBWG7vlNbWvX9aE/fZLUlP2vTcYu+U1SYdFlwHsVfarAO7q6mravjdu3C5J6uzcX0PosKb2L7Av2q8CeMGCBdElAEAFY8AAEIQABoAgBDAABCGAASAIAQwAQQhgAAhCAANAEAIYAIIQwAAQhAAGgCAEMAAEIYABIAgBDABBCGAACEIAA0AQAhgAghDAABCEAAaAIAQwAAQhgAEgiLl7/Y3NXpG0YTePNVnSq7u5bTONxrpGY00SdY0UddVvNNYkNa6uKe5+6OCFIwrgPWFmve4+oyUHG4HRWNdorEmirpGirvqNxpqk5tfFEAQABCGAASBIKwP4xhYeayRGY12jsSaJukaKuuo3GmuSmlxXy8aAAQDVGIIAgCAEMAAEaUgAm1m3mT1rZpmZXVVj/QFmdlda/5iZlQrrvpqWP2tmn2hEPXXW9CUz+28zW2tmPzazKYV1O8zsqfR1X6NqqrOuHjN7pXD8ywrr5pjZz9LXnBbX9c1CTc+Z2RuFdU3pLzO7ycxeNrOnh1hvZnZdqnmtmZ1cWNfMvhqurotSPevM7BEz+1hhXV9a/pSZ9ba4rjPNbFPhsfpaYd0uH/8m1vTlQj1Pp3NpUlrXzL46wsxWpgx4xswW1WjT/PPL3ffoS9JYSc9LmirpA5LWSJo2qM18Sd9Jty+UdFe6PS21P0DSUWk/Y1tU00xJE9LtPxuoKd3fvKc17EFdPZKur7HtJEkvpO8T0+2JraprUPsFkm5qQX/9nqSTJT09xPrZkh6QZJJ+R9Jjze6rOus6feB4kmYN1JXu90maHNRfZ0r61z19/BtZ06C250p6qEV9dbikk9PtD0p6rsZzsennVyOugE+VlLn7C+7+jqQ7JZ0/qM35km5Jt++W9AdmZmn5ne6+zd1flJSl/TW9Jndf6e5b0t1HJX24Acfd47p24ROSVrj7a+7+uqQVkrqD6vqspDsadOwhufu/S3ptF03Ol3Sr5x6VdLCZHa7m9tWwdbn7I+m4UuvOrXr6ayh7cl42sqaWnFeS5O4vufuT6favJP1UUuegZk0/vxoRwJ2Sfl64/7/a+QeptHH37ZI2STqkzm2bVVPRpcp/0w0Yb2a9ZvaomX2yAfWMtK5PpT957jazI0a4bTPrUhqqOUrSQ4XFzeqv4QxVdzP7aqQGn1su6UEzW21m8wLqOc3M1pjZA2Z2fFoW3l9mNkF5iN1TWNySvrJ8SPQkSY8NWtX086ttdzbal5jZ5yTNkPT7hcVT3H2jmU2V9JCZrXP351tU0v2S7nD3bWZ2ufK/HM5q0bHrcaGku919R2FZZH+NWmY2U3kAn1FYfEbqq1+TtMLM1qerxFZ4UvljtdnMZkv6oaRjWnTs4Zwr6SfuXrxabnpfmVmH8tD/c3d/s5H7rkcjroA3SjqicP/DaVnNNmbWJukgSf11btusmmRmfyhpsaTz3H3bwHJ335i+vyBplfLfjo0wbF3u3l+o5buSTql322bWVXChBv2Z2MT+Gs5QdTezr+piZicof/zOd/f+geWFvnpZ0r1qzJBbXdz9TXffnG4vkzTOzCZrFPSXdn1eNaWvzGyc8vC93d1/UKNJ88+vBgxmtykfhD5K7w/gHz+ozRdU/SLcP6fbx6v6RbgX1JgX4eqp6STlLzwcM2j5REkHpNuTJf1MjXtBop66Di/c/mNJj/r7A/8vpvomptuTWlVXanec8hdGrBX9lfZZ0tAvKp2j6hdJHm92X9VZ15HKX884fdDyAyV9sHD7EUndLazr1wceO+Vh9j+p7+p6/JtRU1p/kPJx4gNb1Vfp575V0rd20abp51ejfpjZyl9FfF7S4rTsr5RfWUrSeEnfTyfl45KmFrZdnLZ7VtKsBnbwcDX9m6RfSnoqfd2Xlp8uaV06CddJurRRNdVZ199IeiYdf6Wk4wrbzk19mEm6pJV1pftLJP3toO2a1l/Kr4hekvSu8nG2SyV9XtLn03qT9A+p5nWSZrSor4ar67uSXi+cW71p+dTUT2vSY7y4xXV9sXBuParCL4haj38rakptepS/GF/crtl9dYbyMea1hcdpdqvPLz6KDABB+CQcAAQhgAEgCAEMAEEIYAAIQgADQBACGJIkM1ucZoVam2af+u1h2t9sZhe0qr56pJpeLMyu9UiD9rvEzK5sxL6Aov3+o8iQzOw0SX+kfHaobenTUR8ILmtYZjbWqz8SLUlfdve7QwoCRogrYEj51HyvevoItLu/6u7/J0lm9jUzeyLN1XpjmsWuipmdYmYPp0lTfpRmjJKZLbT351y+s8Z2PWb2L2a2Ks2renVh3efM7PF0JfuPZjY2Ld9sZn9vZmsknVbPD5euYG8xs/8wsw1m9idm9ndprtnl6SOpA/PPDix/3My6auzrxDTp0Fozu9fMJprZ0Wb2ZKHNMQP3d9E3R6djr051HVfPz4J9CwEMSXpQ0hGWT7R+g5kVJya63t1/y90/Kqld+ZVyRQqvpZIucPdTJN0k6a/T6qskneTuJyj/hFEtp0r6lKQTJH3azGaY2W9K+oyk33X3EyXtkHRRan+g8nlZP+bu/1ljf98oDEHcXlh+tPJJjc6TdJukle4+XdJW5R85HbApLb9e0rdq7P9WSV9JP9M6SVd7PvHQJjM7MbW5RNL3humbGyUtSMuvlHTDEP2DfRhDEJDns2OdIunjyieqv8vMrnL3myXNNLO/kDRB+Wfgn1E+Y9uAj0j6qPLZqqR8cu+X0rq1km43sx8qn3mrlhWeJqsxsx8o/4joduWTED2R9tku6eXUfoeqpywcbKghiAfc/V0zW5dqXJ6Wr1M+V8GAOwrfv1ncgZkdJOlgd384LbpF+Ufspfzjx5eY2ZeU//I4VUP0TZqB63RJ3y/8QXHALn4m7KMIYEiS0ljqKkmrUkjNScMGNyj/DPzPzWyJ8nk9ikzSM+5eazjgHOX/EeFcSYvNbLrn80FXHbrGfZN0i7t/tcY+364x7luPgeGV98zsXX//M/jvqfp54EPcHs49kq5WPk/yanfvN7PfUI2+MbMPSXojXd1jP8YQBGRmHzGz4rywJ0raoPfD9tV01VbrXQ/PSjo0vZAnMxtnZseb2RhJR7j7SklfUT7jVUeN7c82s0lm1i7pk5J+IunHki6wfB5YpfVTamzbDJ8pfP+v4gp33yTpdTP7eFr0p5IeTuvelvQjSd+W9L20vmbfeD7v7Itm9um03Kzwf+Ow/+AKGFIejEvN7GDlf/5nkua5+xtm9k+Snpb0C0lPDN7Q3d+x/O1o16U/0duUj50+J+m2tMwkXefubwzeXvnsePcon1P1NnfvlSQzKyv/bwhjlM+k9QXlvxSG84207YCRziE70czWKr9i/myN9XMkfcfy/+DwgvLx3gG3K59C9EFpl33zjPIx7W+nWscp/zdAa0ZYK/ZyzIaGMGbWo3x444vRtUj5uyCU1/Pqbm5/paSD3P0vG1oY9llcAQMNYGb36v13WgB14QoYAILwIhwABCGAASAIAQwAQQhgAAhCAANAkP8H8YAGvEe0fZEAAAAASUVORK5CYII=\n","text/plain":["<Figure size 432x288 with 1 Axes>"]},"metadata":{"tags":[],"needs_background":"light"}}]},{"cell_type":"markdown","metadata":{"id":"baC3dnI0fsYr"},"source":["##### Interest Expense/EBIT"]},{"cell_type":"code","metadata":{"colab":{"base_uri":"https://localhost:8080/","height":333},"id":"Zgri9rJ8fsYr","executionInfo":{"status":"ok","timestamp":1625713110556,"user_tz":240,"elapsed":242,"user":{"displayName":"Tzu-Hua Huang","photoUrl":"https://lh3.googleusercontent.com/a-/AOh14GgbXOdZWwXGVTDuMHbeSme5C5oKAeF6I6ojY6qc=s64","userId":"01458545184256376706"}},"outputId":"b7019071-c45a-4b86-c962-79932d3a0016"},"source":["# check outliers with boxplot\n","sns.boxplot(df1['Interest Expense/EBIT'])\n","plt.show()"],"execution_count":212,"outputs":[{"output_type":"stream","text":["/usr/local/lib/python3.7/dist-packages/seaborn/_decorators.py:43: FutureWarning: Pass the following variable as a keyword arg: x. From version 0.12, the only valid positional argument will be `data`, and passing other arguments without an explicit keyword will result in an error or misinterpretation.\n","  FutureWarning\n"],"name":"stderr"},{"output_type":"display_data","data":{"image/png":"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\n","text/plain":["<Figure size 432x288 with 1 Axes>"]},"metadata":{"tags":[],"needs_background":"light"}}]},{"cell_type":"code","metadata":{"colab":{"base_uri":"https://localhost:8080/"},"id":"aF1KzHnMfsYr","executionInfo":{"status":"ok","timestamp":1625713110557,"user_tz":240,"elapsed":27,"user":{"displayName":"Tzu-Hua Huang","photoUrl":"https://lh3.googleusercontent.com/a-/AOh14GgbXOdZWwXGVTDuMHbeSme5C5oKAeF6I6ojY6qc=s64","userId":"01458545184256376706"}},"outputId":"77783f9d-076f-4da0-d3ae-e066b0c07da2"},"source":["# deal with outliers by winzoring\n","print(df1['Interest Expense/EBIT'].quantile(0.10))\n","print(df1['Interest Expense/EBIT'].quantile(0.94))"],"execution_count":213,"outputs":[{"output_type":"stream","text":["-0.08178241564134169\n","0.9361319435924855\n"],"name":"stdout"}]},{"cell_type":"code","metadata":{"colab":{"base_uri":"https://localhost:8080/"},"id":"ZhCfsulX0VxW","executionInfo":{"status":"ok","timestamp":1625713110558,"user_tz":240,"elapsed":19,"user":{"displayName":"Tzu-Hua Huang","photoUrl":"https://lh3.googleusercontent.com/a-/AOh14GgbXOdZWwXGVTDuMHbeSme5C5oKAeF6I6ojY6qc=s64","userId":"01458545184256376706"}},"outputId":"6bc49839-f5c1-4498-cd43-94cd1f8f49b7"},"source":["# remove the effect of outliers\n","# according to the boxplot, replace 5% of the highest values by winsorizing\n","df1['Interest Expense/EBIT'].clip(lower=df1['Interest Expense/EBIT'].quantile(0.10),\n","                                  upper=df1['Interest Expense/EBIT'].quantile(0.94),\n","                                  axis=0, inplace=True)\n","df1['Interest Expense/EBIT'].describe()"],"execution_count":214,"outputs":[{"output_type":"execute_result","data":{"text/plain":["count    2022.000000\n","mean        0.234694\n","std         0.280338\n","min        -0.081782\n","25%         0.031697\n","50%         0.151023\n","75%         0.364305\n","max         0.936132\n","Name: Interest Expense/EBIT, dtype: float64"]},"metadata":{"tags":[]},"execution_count":214}]},{"cell_type":"code","metadata":{"colab":{"base_uri":"https://localhost:8080/","height":333},"id":"SbpS2tutfsYr","executionInfo":{"status":"ok","timestamp":1625713110723,"user_tz":240,"elapsed":177,"user":{"displayName":"Tzu-Hua Huang","photoUrl":"https://lh3.googleusercontent.com/a-/AOh14GgbXOdZWwXGVTDuMHbeSme5C5oKAeF6I6ojY6qc=s64","userId":"01458545184256376706"}},"outputId":"21b4b927-607d-45a5-90e5-bb13ca475e32"},"source":["# check if outliers exist after winzoring\n","sns.boxplot(df1['Interest Expense/EBIT'])\n","plt.show()"],"execution_count":215,"outputs":[{"output_type":"stream","text":["/usr/local/lib/python3.7/dist-packages/seaborn/_decorators.py:43: FutureWarning: Pass the following variable as a keyword arg: x. From version 0.12, the only valid positional argument will be `data`, and passing other arguments without an explicit keyword will result in an error or misinterpretation.\n","  FutureWarning\n"],"name":"stderr"},{"output_type":"display_data","data":{"image/png":"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\n","text/plain":["<Figure size 432x288 with 1 Axes>"]},"metadata":{"tags":[],"needs_background":"light"}}]},{"cell_type":"markdown","metadata":{"id":"f6JJuAHDdejs"},"source":["## Step 4. Examine the correlation\n","(e) Examine the correlation between *Price/Book, P/E, P/EBIT, P/SALES, RetEarly2020,* and *RetLate2020*. In a text box, explain the correlations you observe.\n","\n","CORRECTION: According to the above process, we will use the ratios as follows: `Book/Price`, `E/P`, `EBIT/P`, `SALES/P`."]},{"cell_type":"code","metadata":{"colab":{"base_uri":"https://localhost:8080/","height":371},"id":"fYtJRjVwdej2","executionInfo":{"status":"ok","timestamp":1625713111426,"user_tz":240,"elapsed":714,"user":{"displayName":"Tzu-Hua Huang","photoUrl":"https://lh3.googleusercontent.com/a-/AOh14GgbXOdZWwXGVTDuMHbeSme5C5oKAeF6I6ojY6qc=s64","userId":"01458545184256376706"}},"outputId":"c192b0b3-b8f7-4aca-e875-5548b2926ffe"},"source":["# create correlation matric to check the correlation between the variables\n","corr_matrix = df1.corr()\n","sns.heatmap(corr_matrix, annot=True)\n","plt.show()"],"execution_count":216,"outputs":[{"output_type":"display_data","data":{"image/png":"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\n","text/plain":["<Figure size 432x288 with 2 Axes>"]},"metadata":{"tags":[],"needs_background":"light"}}]},{"cell_type":"markdown","metadata":{"id":"5kHneXkaVjfR"},"source":["#### Correlation Matrix Interpretation\n","`Cash/CL`: Cash/Current Liabilities\n","\n","`Long-term Debt/TA`: Long-term Debt/Total Assets\n","\n","\n","1. From the correlation matrix, in general, the correlation among the market ratios, are higher than the correlation between any of the market ratios and `RetEarly2020` and `RetLate2020`. \n","2. The correlation between `Quick Ratio` and `Cash/CL` is the highest among the X variables. Both of them use `Current Liabilities` as denominator, since liquid assets in `Quick Ratio` includes cash on hand or assets other than cash that can be quickly converted into cash, the high correlation makes sense.\n","3. The correlation between `Long-term Debt/TA` and `Cash/Interest Expense` have the highest negative correlation. We may infer that the firms with lower long-term debt have more cash to cover\n","immediate interest expenses. \n","4. Compared to `RetLate2020`, the correlations between each of market ratios and `RetEarly2020` are stronger. Since the ratios are derived from 2019, the connection might be stronger when it compares to the data closer to 2019, which is `RetEarly2020` in this case.\n","5. The correaltions between market ratios and `RetEarly2020` and `RetLate2020` are generally negative, which indicates that they’re negatively correlated. Lower market ratios are related to higher `RetEarly2020` and `RetLate2020`."]},{"cell_type":"markdown","metadata":{"id":"24Ei8RhRdjQT"},"source":["## Step 5. Estimate linear regression models"]},{"cell_type":"markdown","metadata":{"id":"EZGqvpprjqF_"},"source":["### 5.a RetEarly2020"]},{"cell_type":"code","metadata":{"id":"6CHzgtBTjqF_","executionInfo":{"status":"ok","timestamp":1625713111427,"user_tz":240,"elapsed":16,"user":{"displayName":"Tzu-Hua Huang","photoUrl":"https://lh3.googleusercontent.com/a-/AOh14GgbXOdZWwXGVTDuMHbeSme5C5oKAeF6I6ojY6qc=s64","userId":"01458545184256376706"}}},"source":["# define the independent variables (X)\n","x1 = df1[['Cash/TA', 'Cash/CL', 'Cash/Interest Expense', 'Quick Ratio', 'Long-term Debt/TA',\n","          'Fixed Assets/TA', 'Sales per Employee', 'Interest Expense/EBIT']]\n","\n","# define the dependent variable (Y)\n","y1 = df1['RetEarly2020']"],"execution_count":217,"outputs":[]},{"cell_type":"code","metadata":{"colab":{"base_uri":"https://localhost:8080/"},"id":"bxD9J9OUjqGA","executionInfo":{"status":"ok","timestamp":1625713111428,"user_tz":240,"elapsed":16,"user":{"displayName":"Tzu-Hua Huang","photoUrl":"https://lh3.googleusercontent.com/a-/AOh14GgbXOdZWwXGVTDuMHbeSme5C5oKAeF6I6ojY6qc=s64","userId":"01458545184256376706"}},"outputId":"c70141b6-24e3-43e9-9cf7-5a8148878bfb"},"source":["# run multiple regression with constant\n","# use statsmodels\n","# add a constant\n","x1 = sm.add_constant(x1) \n","model1 = sm.OLS(y1, x1).fit()\n","\n","# print out regression summary statistics\n","print_model1 = model1.summary()\n","print(print_model1)"],"execution_count":218,"outputs":[{"output_type":"stream","text":["                            OLS Regression Results                            \n","==============================================================================\n","Dep. Variable:           RetEarly2020   R-squared:                       0.063\n","Model:                            OLS   Adj. R-squared:                  0.059\n","Method:                 Least Squares   F-statistic:                     16.97\n","Date:                Thu, 08 Jul 2021   Prob (F-statistic):           1.33e-24\n","Time:                        02:58:31   Log-Likelihood:                -550.26\n","No. Observations:                2022   AIC:                             1119.\n","Df Residuals:                    2013   BIC:                             1169.\n","Df Model:                           8                                         \n","Covariance Type:            nonrobust                                         \n","=========================================================================================\n","                            coef    std err          t      P>|t|      [0.025      0.975]\n","-----------------------------------------------------------------------------------------\n","const                    -0.2541      0.027     -9.502      0.000      -0.307      -0.202\n","Cash/TA                   0.2679      0.080      3.335      0.001       0.110       0.425\n","Cash/CL                   0.0459      0.019      2.424      0.015       0.009       0.083\n","Cash/Interest Expense  5.875e-05      0.001      0.079      0.937      -0.001       0.002\n","Quick Ratio              -0.0245      0.012     -2.042      0.041      -0.048      -0.001\n","Long-term Debt/TA        -0.0479      0.048     -0.998      0.318      -0.142       0.046\n","Fixed Assets/TA          -0.0834      0.034     -2.443      0.015      -0.150      -0.016\n","Sales per Employee       -0.0355      0.014     -2.546      0.011      -0.063      -0.008\n","Interest Expense/EBIT    -0.0884      0.030     -2.940      0.003      -0.147      -0.029\n","==============================================================================\n","Omnibus:                     3201.716   Durbin-Watson:                   1.836\n","Prob(Omnibus):                  0.000   Jarque-Bera (JB):          3183096.997\n","Skew:                           9.643   Prob(JB):                         0.00\n","Kurtosis:                     196.416   Cond. No.                         260.\n","==============================================================================\n","\n","Warnings:\n","[1] Standard Errors assume that the covariance matrix of the errors is correctly specified.\n"],"name":"stdout"}]},{"cell_type":"markdown","metadata":{"id":"vymW3cD-2vaD"},"source":["**Interpretation**\n","\n","`R-squared: 0.063`\n","\n","Around 6% of the differences or the variation in observed early 2020 return for companies is explained by the financial ratios, indicating 6% of the variation which varies across all these companies is explained by their financial ratios.\n","\n","`Adj. R-squared: 0.059`\n","\n","The adjusted R-squared considers and tests different independent variables against the model. Compared to the model with additional input variables, decreased adjusted R-squared indicates that the additional input variables are not adding value to the model."]},{"cell_type":"code","metadata":{"colab":{"base_uri":"https://localhost:8080/","height":328},"id":"iC46F-02By27","executionInfo":{"status":"ok","timestamp":1625713111429,"user_tz":240,"elapsed":13,"user":{"displayName":"Tzu-Hua Huang","photoUrl":"https://lh3.googleusercontent.com/a-/AOh14GgbXOdZWwXGVTDuMHbeSme5C5oKAeF6I6ojY6qc=s64","userId":"01458545184256376706"}},"outputId":"65c23cf0-f0bf-4ea2-a201-fe38c02e0b7f"},"source":["df1_table = pd.read_html(print_model1.tables[1].as_html(),header=0,index_col=0)[0]\n","df1_table"],"execution_count":219,"outputs":[{"output_type":"execute_result","data":{"text/html":["<div>\n","<style scoped>\n","    .dataframe tbody tr th:only-of-type {\n","        vertical-align: middle;\n","    }\n","\n","    .dataframe tbody tr th {\n","        vertical-align: top;\n","    }\n","\n","    .dataframe thead th {\n","        text-align: right;\n","    }\n","</style>\n","<table border=\"1\" class=\"dataframe\">\n","  <thead>\n","    <tr style=\"text-align: right;\">\n","      <th></th>\n","      <th>coef</th>\n","      <th>std err</th>\n","      <th>t</th>\n","      <th>P&gt;|t|</th>\n","      <th>[0.025</th>\n","      <th>0.975]</th>\n","    </tr>\n","  </thead>\n","  <tbody>\n","    <tr>\n","      <th>const</th>\n","      <td>-0.254100</td>\n","      <td>0.027</td>\n","      <td>-9.502</td>\n","      <td>0.000</td>\n","      <td>-0.307</td>\n","      <td>-0.202</td>\n","    </tr>\n","    <tr>\n","      <th>Cash/TA</th>\n","      <td>0.267900</td>\n","      <td>0.080</td>\n","      <td>3.335</td>\n","      <td>0.001</td>\n","      <td>0.110</td>\n","      <td>0.425</td>\n","    </tr>\n","    <tr>\n","      <th>Cash/CL</th>\n","      <td>0.045900</td>\n","      <td>0.019</td>\n","      <td>2.424</td>\n","      <td>0.015</td>\n","      <td>0.009</td>\n","      <td>0.083</td>\n","    </tr>\n","    <tr>\n","      <th>Cash/Interest Expense</th>\n","      <td>0.000059</td>\n","      <td>0.001</td>\n","      <td>0.079</td>\n","      <td>0.937</td>\n","      <td>-0.001</td>\n","      <td>0.002</td>\n","    </tr>\n","    <tr>\n","      <th>Quick Ratio</th>\n","      <td>-0.024500</td>\n","      <td>0.012</td>\n","      <td>-2.042</td>\n","      <td>0.041</td>\n","      <td>-0.048</td>\n","      <td>-0.001</td>\n","    </tr>\n","    <tr>\n","      <th>Long-term Debt/TA</th>\n","      <td>-0.047900</td>\n","      <td>0.048</td>\n","      <td>-0.998</td>\n","      <td>0.318</td>\n","      <td>-0.142</td>\n","      <td>0.046</td>\n","    </tr>\n","    <tr>\n","      <th>Fixed Assets/TA</th>\n","      <td>-0.083400</td>\n","      <td>0.034</td>\n","      <td>-2.443</td>\n","      <td>0.015</td>\n","      <td>-0.150</td>\n","      <td>-0.016</td>\n","    </tr>\n","    <tr>\n","      <th>Sales per Employee</th>\n","      <td>-0.035500</td>\n","      <td>0.014</td>\n","      <td>-2.546</td>\n","      <td>0.011</td>\n","      <td>-0.063</td>\n","      <td>-0.008</td>\n","    </tr>\n","    <tr>\n","      <th>Interest Expense/EBIT</th>\n","      <td>-0.088400</td>\n","      <td>0.030</td>\n","      <td>-2.940</td>\n","      <td>0.003</td>\n","      <td>-0.147</td>\n","      <td>-0.029</td>\n","    </tr>\n","  </tbody>\n","</table>\n","</div>"],"text/plain":["                           coef  std err      t  P>|t|  [0.025  0.975]\n","const                 -0.254100    0.027 -9.502  0.000  -0.307  -0.202\n","Cash/TA                0.267900    0.080  3.335  0.001   0.110   0.425\n","Cash/CL                0.045900    0.019  2.424  0.015   0.009   0.083\n","Cash/Interest Expense  0.000059    0.001  0.079  0.937  -0.001   0.002\n","Quick Ratio           -0.024500    0.012 -2.042  0.041  -0.048  -0.001\n","Long-term Debt/TA     -0.047900    0.048 -0.998  0.318  -0.142   0.046\n","Fixed Assets/TA       -0.083400    0.034 -2.443  0.015  -0.150  -0.016\n","Sales per Employee    -0.035500    0.014 -2.546  0.011  -0.063  -0.008\n","Interest Expense/EBIT -0.088400    0.030 -2.940  0.003  -0.147  -0.029"]},"metadata":{"tags":[]},"execution_count":219}]},{"cell_type":"markdown","metadata":{"id":"Xue4s1yp5DWJ"},"source":["**t-stats**\n","According to the coefficients in the above summary table:\n","- **Cash/TA**\n","- **Cash/CL**\n","- **Quick Ratio**\n","- **Fixed Assets/TA**\n","- **Sales per Employee**\n","- **Interest Expense/EBIT**\n","  - The lager the magnitude of the t value, the more the evidence against the null hypothesis, indicating that the more likely there is a significant difference. \n","  - The p-values for the above variables are less than 0.05. We can reject the null hypothesis, suggesting that the variable has some effect with a 95% probability of being correct, assuming the model is specified correctly. \n","  - We can conclude that the data provide enough evidence to reject the null hypothesis, suggesting that each variable is statistically significant.\n","\n","- **Cash/Interest Expense**\n","- **Long-term Debt/TA**\n","  - The t-values are close to 0. The greater the t-values, the more likely there isn't a significant difference. Their p-values are greater than 0.05. We fail to reject the null hypothesis."]},{"cell_type":"markdown","metadata":{"id":"krDO9MRljqGA"},"source":["### 5.b RetLate2020"]},{"cell_type":"code","metadata":{"id":"ZiJNbl8NjqGA","executionInfo":{"status":"ok","timestamp":1625713111430,"user_tz":240,"elapsed":12,"user":{"displayName":"Tzu-Hua Huang","photoUrl":"https://lh3.googleusercontent.com/a-/AOh14GgbXOdZWwXGVTDuMHbeSme5C5oKAeF6I6ojY6qc=s64","userId":"01458545184256376706"}}},"source":["# define the independent variables (X)\n","x2 = df1[['Cash/TA', 'Cash/CL', 'Cash/Interest Expense', 'Quick Ratio', 'Long-term Debt/TA',\n","          'Fixed Assets/TA', 'Sales per Employee', 'Interest Expense/EBIT']]\n","\n","# define the dependent variable (Y)\n","y2 = df1['RetLate2020']"],"execution_count":220,"outputs":[]},{"cell_type":"code","metadata":{"colab":{"base_uri":"https://localhost:8080/"},"id":"N2QQy2iUjqGA","executionInfo":{"status":"ok","timestamp":1625713111620,"user_tz":240,"elapsed":201,"user":{"displayName":"Tzu-Hua Huang","photoUrl":"https://lh3.googleusercontent.com/a-/AOh14GgbXOdZWwXGVTDuMHbeSme5C5oKAeF6I6ojY6qc=s64","userId":"01458545184256376706"}},"outputId":"57f8dc99-2666-40a1-b7b9-bab495e05304"},"source":["# run multiple regression with constant\n","# use statsmodels\n","# add a constant\n","x2 = sm.add_constant(x2) \n","model2 = sm.OLS(y2, x2).fit()\n","\n","# print out regression summary statistics\n","print_model2 = mode2.summary()\n","print(print_model)"],"execution_count":221,"outputs":[{"output_type":"stream","text":["                            OLS Regression Results                            \n","==============================================================================\n","Dep. Variable:            RetLate2020   R-squared:                       0.040\n","Model:                            OLS   Adj. R-squared:                  0.036\n","Method:                 Least Squares   F-statistic:                     10.45\n","Date:                Thu, 08 Jul 2021   Prob (F-statistic):           1.88e-14\n","Time:                        02:01:53   Log-Likelihood:                -3008.4\n","No. Observations:                2022   AIC:                             6035.\n","Df Residuals:                    2013   BIC:                             6085.\n","Df Model:                           8                                         \n","Covariance Type:            nonrobust                                         \n","=========================================================================================\n","                            coef    std err          t      P>|t|      [0.025      0.975]\n","-----------------------------------------------------------------------------------------\n","const                     0.8345      0.090      9.253      0.000       0.658       1.011\n","Cash/TA                   1.4419      0.271      5.321      0.000       0.911       1.973\n","Cash/CL                  -0.0968      0.064     -1.516      0.130      -0.222       0.028\n","Cash/Interest Expense    -0.0004      0.003     -0.149      0.881      -0.005       0.005\n","Quick Ratio              -0.0930      0.040     -2.298      0.022      -0.172      -0.014\n","Long-term Debt/TA         0.3867      0.162      2.390      0.017       0.069       0.704\n","Fixed Assets/TA           0.1161      0.115      1.009      0.313      -0.110       0.342\n","Sales per Employee       -0.0724      0.047     -1.538      0.124      -0.165       0.020\n","Interest Expense/EBIT    -0.0209      0.101     -0.206      0.837      -0.220       0.178\n","==============================================================================\n","Omnibus:                     2196.425   Durbin-Watson:                   1.996\n","Prob(Omnibus):                  0.000   Jarque-Bera (JB):           272767.904\n","Skew:                           5.199   Prob(JB):                         0.00\n","Kurtosis:                      58.942   Cond. No.                         260.\n","==============================================================================\n","\n","Warnings:\n","[1] Standard Errors assume that the covariance matrix of the errors is correctly specified.\n"],"name":"stdout"}]},{"cell_type":"markdown","metadata":{"id":"z2kcm7zJdej3"},"source":["**Interpretation**\n","\n","`R-squared: 0.040`\n","\n","Around 4% of the differences or the variation in observed late 2020 return for companies is explained by the financial ratios, indicating 4% of the variation which varies across all these companies is explained by their financial ratios.\n","\n","`Adj. R-squared: 0.036`\n","\n","The adjusted R-squared considers and tests different independent variables against the model. Compared to the model with additional input variables, decreased adjusted R-squared indicates that the additional input variables are not adding value to the model."]},{"cell_type":"code","metadata":{"colab":{"base_uri":"https://localhost:8080/","height":328},"id":"QOaNK5ga9KKC","executionInfo":{"status":"ok","timestamp":1625713111621,"user_tz":240,"elapsed":20,"user":{"displayName":"Tzu-Hua Huang","photoUrl":"https://lh3.googleusercontent.com/a-/AOh14GgbXOdZWwXGVTDuMHbeSme5C5oKAeF6I6ojY6qc=s64","userId":"01458545184256376706"}},"outputId":"bdf1dcec-48e9-49a0-a5b0-1247cf9ba2b5"},"source":["df2_table = pd.read_html(print_model2.tables[1].as_html(),header=0,index_col=0)[0]\n","df2_table"],"execution_count":222,"outputs":[{"output_type":"execute_result","data":{"text/html":["<div>\n","<style scoped>\n","    .dataframe tbody tr th:only-of-type {\n","        vertical-align: middle;\n","    }\n","\n","    .dataframe tbody tr th {\n","        vertical-align: top;\n","    }\n","\n","    .dataframe thead th {\n","        text-align: right;\n","    }\n","</style>\n","<table border=\"1\" class=\"dataframe\">\n","  <thead>\n","    <tr style=\"text-align: right;\">\n","      <th></th>\n","      <th>coef</th>\n","      <th>std err</th>\n","      <th>t</th>\n","      <th>P&gt;|t|</th>\n","      <th>[0.025</th>\n","      <th>0.975]</th>\n","    </tr>\n","  </thead>\n","  <tbody>\n","    <tr>\n","      <th>const</th>\n","      <td>0.8345</td>\n","      <td>0.090</td>\n","      <td>9.253</td>\n","      <td>0.000</td>\n","      <td>0.658</td>\n","      <td>1.011</td>\n","    </tr>\n","    <tr>\n","      <th>Cash/TA</th>\n","      <td>1.4419</td>\n","      <td>0.271</td>\n","      <td>5.321</td>\n","      <td>0.000</td>\n","      <td>0.911</td>\n","      <td>1.973</td>\n","    </tr>\n","    <tr>\n","      <th>Cash/CL</th>\n","      <td>-0.0968</td>\n","      <td>0.064</td>\n","      <td>-1.516</td>\n","      <td>0.130</td>\n","      <td>-0.222</td>\n","      <td>0.028</td>\n","    </tr>\n","    <tr>\n","      <th>Cash/Interest Expense</th>\n","      <td>-0.0004</td>\n","      <td>0.003</td>\n","      <td>-0.149</td>\n","      <td>0.881</td>\n","      <td>-0.005</td>\n","      <td>0.005</td>\n","    </tr>\n","    <tr>\n","      <th>Quick Ratio</th>\n","      <td>-0.0930</td>\n","      <td>0.040</td>\n","      <td>-2.298</td>\n","      <td>0.022</td>\n","      <td>-0.172</td>\n","      <td>-0.014</td>\n","    </tr>\n","    <tr>\n","      <th>Long-term Debt/TA</th>\n","      <td>0.3867</td>\n","      <td>0.162</td>\n","      <td>2.390</td>\n","      <td>0.017</td>\n","      <td>0.069</td>\n","      <td>0.704</td>\n","    </tr>\n","    <tr>\n","      <th>Fixed Assets/TA</th>\n","      <td>0.1161</td>\n","      <td>0.115</td>\n","      <td>1.009</td>\n","      <td>0.313</td>\n","      <td>-0.110</td>\n","      <td>0.342</td>\n","    </tr>\n","    <tr>\n","      <th>Sales per Employee</th>\n","      <td>-0.0724</td>\n","      <td>0.047</td>\n","      <td>-1.538</td>\n","      <td>0.124</td>\n","      <td>-0.165</td>\n","      <td>0.020</td>\n","    </tr>\n","    <tr>\n","      <th>Interest Expense/EBIT</th>\n","      <td>-0.0209</td>\n","      <td>0.101</td>\n","      <td>-0.206</td>\n","      <td>0.837</td>\n","      <td>-0.220</td>\n","      <td>0.178</td>\n","    </tr>\n","  </tbody>\n","</table>\n","</div>"],"text/plain":["                         coef  std err      t  P>|t|  [0.025  0.975]\n","const                  0.8345    0.090  9.253  0.000   0.658   1.011\n","Cash/TA                1.4419    0.271  5.321  0.000   0.911   1.973\n","Cash/CL               -0.0968    0.064 -1.516  0.130  -0.222   0.028\n","Cash/Interest Expense -0.0004    0.003 -0.149  0.881  -0.005   0.005\n","Quick Ratio           -0.0930    0.040 -2.298  0.022  -0.172  -0.014\n","Long-term Debt/TA      0.3867    0.162  2.390  0.017   0.069   0.704\n","Fixed Assets/TA        0.1161    0.115  1.009  0.313  -0.110   0.342\n","Sales per Employee    -0.0724    0.047 -1.538  0.124  -0.165   0.020\n","Interest Expense/EBIT -0.0209    0.101 -0.206  0.837  -0.220   0.178"]},"metadata":{"tags":[]},"execution_count":222}]},{"cell_type":"markdown","metadata":{"id":"PwkGNu5h9Keu"},"source":["**t-stats**\n","According to the coefficients in the above summary table:\n","- **Cash/TA**\n","- **Quick Ratio**\n","- **Long-term Debt/TA**\n","  - The lager the magnitude of the t value, the more the evidence against the null hypothesis, indicating that the more likely there is a significant difference. \n","  - The p-values for the above variables are less than 0.05. We can reject the null hypothesis, suggesting that the variable has some effect with a 95% probability of being correct, assuming the model is specified correctly. \n","  - We can conclude that the data provide enough evidence to reject the null hypothesis, suggesting that each variable is statistically significant.\n","\n","- **Cash/CL**\n","- **Cash/Interest Expense**\n","- **Fixed Assets/TA**\n","- **Sales per Employee**\n","- **Interest Expense/EBIT**\n","  - The t-values are close to 0. The greater the t-values, the more likely there isn't a significant difference. Their p-values are greater than 0.05. We fail to reject the null hypothesis."]},{"cell_type":"markdown","metadata":{"id":"Y_OEwdbk-_W6"},"source":["According to the coefficients in the above summary table:\n","- **Cash/TA**: The firms with more cash at the end of 2019 have\n","higher stock returns in 2020 after Covid shock.\n","- **Quick Ratio**: In contrast, firms with more liquid assets to cover current\n","liabilities at end of 2019 have lower stock returns in 2020 after\n","Covid shock.\n","- **Long-term Debt/TA**: The firms higher long-term\n","debt at end of 2019 have higher stock returns in 2020 after Covid\n","shock.\n","\n","However, if we consider the R-Squared/Adj. R-Squared, only around 4% of the variation in observed late 2020 return for companies can be explained by the financial ratios. The proportion of the data fit the regression model is quite low. We may consider other indicators as explanatory variables and look for higher explanatory power."]},{"cell_type":"markdown","metadata":{"id":"2yX-chLC2qtg"},"source":["**RetEarly2020 v.s. RetLate2020**\n","\n","If we compare the Adj. R-squared of `RetEarly2020` versus `RetLate2020`, the Adj. R-squared of `RetEarly2020` is higher than `RetLate2020`. For the model of `RetEarly2020`, the higher Adj. R-squared means more variation in observed 2020 return can be explained by the financials, indicating a better fit for the model. We may infer that during 2020, the pandemic caused volatility and instability to the market. Compared to `RetLate2020`, the model for `RetLate2020` covering longer unstable period, thus has lower explanatory power."]},{"cell_type":"markdown","metadata":{"id":"LzkMDqNgG6eI"},"source":["# 2.Does Industry or Sector Explain Differences in Stock Returns for Early (Jan-Mar) and Late (Apr-Dec) 2020? "]},{"cell_type":"markdown","metadata":{"id":"y0yj-r9X-gAk"},"source":["## Step 1. Load and merge data \n","Download and load the data from `temp1.csv`, `tempGICS.csv`, `tempSIC.csv` into Pandas dataframes. Merge the datasets based on each stock's ticker symbol (TIC)."]},{"cell_type":"code","metadata":{"colab":{"base_uri":"https://localhost:8080/"},"id":"TOSpgbaX-lwV","executionInfo":{"status":"ok","timestamp":1625713111622,"user_tz":240,"elapsed":18,"user":{"displayName":"Tzu-Hua Huang","photoUrl":"https://lh3.googleusercontent.com/a-/AOh14GgbXOdZWwXGVTDuMHbeSme5C5oKAeF6I6ojY6qc=s64","userId":"01458545184256376706"}},"outputId":"a4512244-dd4e-4902-f8ae-2f03dc182a21"},"source":["# # load data\n","# temp1 = pd.read_csv('temp1.csv')\n","\n","# # check data\n","# temp1[temp1.isnull().any(axis=1)]\n","\n","# # drop the rows with errors\n","# temp1.drop(temp1.loc[temp1['RetEarly2020'] == '#VALUE!'].index, inplace=True)\n","\n","# # convert datatype of RetEarly2020 from object to float\n","# temp1['RetEarly2020'] = pd.to_numeric(temp1['RetEarly2020'])\n","temp1.info()"],"execution_count":223,"outputs":[{"output_type":"stream","text":["<class 'pandas.core.frame.DataFrame'>\n","Int64Index: 2711 entries, 0 to 2715\n","Data columns (total 3 columns):\n"," #   Column        Non-Null Count  Dtype  \n","---  ------        --------------  -----  \n"," 0   TICKER        2711 non-null   object \n"," 1   RetEarly2020  2711 non-null   float64\n"," 2   RetLate2020   2711 non-null   float64\n","dtypes: float64(2), object(1)\n","memory usage: 84.7+ KB\n"],"name":"stdout"}]},{"cell_type":"code","metadata":{"colab":{"base_uri":"https://localhost:8080/","height":142},"id":"nqXr41X6-tjG","executionInfo":{"status":"ok","timestamp":1625713111737,"user_tz":240,"elapsed":126,"user":{"displayName":"Tzu-Hua Huang","photoUrl":"https://lh3.googleusercontent.com/a-/AOh14GgbXOdZWwXGVTDuMHbeSme5C5oKAeF6I6ojY6qc=s64","userId":"01458545184256376706"}},"outputId":"17f09461-0bf4-4e68-8b71-986b91eb54e4"},"source":["# load data\n","tempGICS = pd.read_csv('tempGICS.csv')\n","tempGICS.head(3)"],"execution_count":224,"outputs":[{"output_type":"execute_result","data":{"text/html":["<div>\n","<style scoped>\n","    .dataframe tbody tr th:only-of-type {\n","        vertical-align: middle;\n","    }\n","\n","    .dataframe tbody tr th {\n","        vertical-align: top;\n","    }\n","\n","    .dataframe thead th {\n","        text-align: right;\n","    }\n","</style>\n","<table border=\"1\" class=\"dataframe\">\n","  <thead>\n","    <tr style=\"text-align: right;\">\n","      <th></th>\n","      <th>gvkey</th>\n","      <th>datadate</th>\n","      <th>fyear</th>\n","      <th>indfmt</th>\n","      <th>consol</th>\n","      <th>popsrc</th>\n","      <th>datafmt</th>\n","      <th>tic</th>\n","      <th>curcd</th>\n","      <th>costat</th>\n","      <th>ggroup</th>\n","    </tr>\n","  </thead>\n","  <tbody>\n","    <tr>\n","      <th>0</th>\n","      <td>1004</td>\n","      <td>20200531</td>\n","      <td>2019.0</td>\n","      <td>INDL</td>\n","      <td>C</td>\n","      <td>D</td>\n","      <td>STD</td>\n","      <td>AIR</td>\n","      <td>USD</td>\n","      <td>A</td>\n","      <td>2010</td>\n","    </tr>\n","    <tr>\n","      <th>1</th>\n","      <td>1045</td>\n","      <td>20191231</td>\n","      <td>2019.0</td>\n","      <td>INDL</td>\n","      <td>C</td>\n","      <td>D</td>\n","      <td>STD</td>\n","      <td>AAL</td>\n","      <td>USD</td>\n","      <td>A</td>\n","      <td>2030</td>\n","    </tr>\n","    <tr>\n","      <th>2</th>\n","      <td>1050</td>\n","      <td>20191231</td>\n","      <td>2019.0</td>\n","      <td>INDL</td>\n","      <td>C</td>\n","      <td>D</td>\n","      <td>STD</td>\n","      <td>CECE</td>\n","      <td>USD</td>\n","      <td>A</td>\n","      <td>2020</td>\n","    </tr>\n","  </tbody>\n","</table>\n","</div>"],"text/plain":["   gvkey  datadate   fyear indfmt consol  ... datafmt   tic curcd costat ggroup\n","0   1004  20200531  2019.0   INDL      C  ...     STD   AIR   USD      A   2010\n","1   1045  20191231  2019.0   INDL      C  ...     STD   AAL   USD      A   2030\n","2   1050  20191231  2019.0   INDL      C  ...     STD  CECE   USD      A   2020\n","\n","[3 rows x 11 columns]"]},"metadata":{"tags":[]},"execution_count":224}]},{"cell_type":"code","metadata":{"colab":{"base_uri":"https://localhost:8080/","height":142},"id":"djw24F6Rj5gO","executionInfo":{"status":"ok","timestamp":1625713111762,"user_tz":240,"elapsed":49,"user":{"displayName":"Tzu-Hua Huang","photoUrl":"https://lh3.googleusercontent.com/a-/AOh14GgbXOdZWwXGVTDuMHbeSme5C5oKAeF6I6ojY6qc=s64","userId":"01458545184256376706"}},"outputId":"346c2974-873c-4d30-f5d4-d91731444121"},"source":["# load data\n","tempNAICS = pd.read_csv('tempNAICS.csv')\n","tempNAICS.head(3)"],"execution_count":225,"outputs":[{"output_type":"execute_result","data":{"text/html":["<div>\n","<style scoped>\n","    .dataframe tbody tr th:only-of-type {\n","        vertical-align: middle;\n","    }\n","\n","    .dataframe tbody tr th {\n","        vertical-align: top;\n","    }\n","\n","    .dataframe thead th {\n","        text-align: right;\n","    }\n","</style>\n","<table border=\"1\" class=\"dataframe\">\n","  <thead>\n","    <tr style=\"text-align: right;\">\n","      <th></th>\n","      <th>gvkey</th>\n","      <th>datadate</th>\n","      <th>fyear</th>\n","      <th>indfmt</th>\n","      <th>consol</th>\n","      <th>popsrc</th>\n","      <th>datafmt</th>\n","      <th>tic</th>\n","      <th>curcd</th>\n","      <th>costat</th>\n","      <th>naics</th>\n","    </tr>\n","  </thead>\n","  <tbody>\n","    <tr>\n","      <th>0</th>\n","      <td>1004</td>\n","      <td>20200531</td>\n","      <td>2019.0</td>\n","      <td>INDL</td>\n","      <td>C</td>\n","      <td>D</td>\n","      <td>STD</td>\n","      <td>AIR</td>\n","      <td>USD</td>\n","      <td>A</td>\n","      <td>423860</td>\n","    </tr>\n","    <tr>\n","      <th>1</th>\n","      <td>1045</td>\n","      <td>20191231</td>\n","      <td>2019.0</td>\n","      <td>INDL</td>\n","      <td>C</td>\n","      <td>D</td>\n","      <td>STD</td>\n","      <td>AAL</td>\n","      <td>USD</td>\n","      <td>A</td>\n","      <td>481111</td>\n","    </tr>\n","    <tr>\n","      <th>2</th>\n","      <td>1050</td>\n","      <td>20191231</td>\n","      <td>2019.0</td>\n","      <td>INDL</td>\n","      <td>C</td>\n","      <td>D</td>\n","      <td>STD</td>\n","      <td>CECE</td>\n","      <td>USD</td>\n","      <td>A</td>\n","      <td>333413</td>\n","    </tr>\n","  </tbody>\n","</table>\n","</div>"],"text/plain":["   gvkey  datadate   fyear indfmt consol  ... datafmt   tic curcd costat   naics\n","0   1004  20200531  2019.0   INDL      C  ...     STD   AIR   USD      A  423860\n","1   1045  20191231  2019.0   INDL      C  ...     STD   AAL   USD      A  481111\n","2   1050  20191231  2019.0   INDL      C  ...     STD  CECE   USD      A  333413\n","\n","[3 rows x 11 columns]"]},"metadata":{"tags":[]},"execution_count":225}]},{"cell_type":"code","metadata":{"colab":{"base_uri":"https://localhost:8080/","height":224},"id":"p8Ghpf4v-6jl","executionInfo":{"status":"ok","timestamp":1625713111764,"user_tz":240,"elapsed":44,"user":{"displayName":"Tzu-Hua Huang","photoUrl":"https://lh3.googleusercontent.com/a-/AOh14GgbXOdZWwXGVTDuMHbeSme5C5oKAeF6I6ojY6qc=s64","userId":"01458545184256376706"}},"outputId":"5b9d8ba2-0116-4340-b26b-8bf6375b30ad"},"source":["# merge data\n","temp = temp1.merge(tempGICS, left_on='TICKER', right_on='tic')\n","df = temp.merge(tempNAICS, left_on='TICKER', right_on='tic')\n","df.head()"],"execution_count":226,"outputs":[{"output_type":"execute_result","data":{"text/html":["<div>\n","<style scoped>\n","    .dataframe tbody tr th:only-of-type {\n","        vertical-align: middle;\n","    }\n","\n","    .dataframe tbody tr th {\n","        vertical-align: top;\n","    }\n","\n","    .dataframe thead th {\n","        text-align: right;\n","    }\n","</style>\n","<table border=\"1\" class=\"dataframe\">\n","  <thead>\n","    <tr style=\"text-align: right;\">\n","      <th></th>\n","      <th>TICKER</th>\n","      <th>RetEarly2020</th>\n","      <th>RetLate2020</th>\n","      <th>gvkey_x</th>\n","      <th>datadate_x</th>\n","      <th>fyear_x</th>\n","      <th>indfmt_x</th>\n","      <th>consol_x</th>\n","      <th>popsrc_x</th>\n","      <th>datafmt_x</th>\n","      <th>tic_x</th>\n","      <th>curcd_x</th>\n","      <th>costat_x</th>\n","      <th>ggroup</th>\n","      <th>gvkey_y</th>\n","      <th>datadate_y</th>\n","      <th>fyear_y</th>\n","      <th>indfmt_y</th>\n","      <th>consol_y</th>\n","      <th>popsrc_y</th>\n","      <th>datafmt_y</th>\n","      <th>tic_y</th>\n","      <th>curcd_y</th>\n","      <th>costat_y</th>\n","      <th>naics</th>\n","    </tr>\n","  </thead>\n","  <tbody>\n","    <tr>\n","      <th>0</th>\n","      <td>JJSF</td>\n","      <td>-0.340234</td>\n","      <td>0.300340</td>\n","      <td>12825</td>\n","      <td>20190930</td>\n","      <td>2019.0</td>\n","      <td>INDL</td>\n","      <td>C</td>\n","      <td>D</td>\n","      <td>STD</td>\n","      <td>JJSF</td>\n","      <td>USD</td>\n","      <td>A</td>\n","      <td>3020</td>\n","      <td>12825</td>\n","      <td>20190930</td>\n","      <td>2019.0</td>\n","      <td>INDL</td>\n","      <td>C</td>\n","      <td>D</td>\n","      <td>STD</td>\n","      <td>JJSF</td>\n","      <td>USD</td>\n","      <td>A</td>\n","      <td>311812</td>\n","    </tr>\n","    <tr>\n","      <th>1</th>\n","      <td>PLXS</td>\n","      <td>-0.290876</td>\n","      <td>0.433469</td>\n","      <td>12945</td>\n","      <td>20190930</td>\n","      <td>2019.0</td>\n","      <td>INDL</td>\n","      <td>C</td>\n","      <td>D</td>\n","      <td>STD</td>\n","      <td>PLXS</td>\n","      <td>USD</td>\n","      <td>A</td>\n","      <td>4520</td>\n","      <td>12945</td>\n","      <td>20190930</td>\n","      <td>2019.0</td>\n","      <td>INDL</td>\n","      <td>C</td>\n","      <td>D</td>\n","      <td>STD</td>\n","      <td>PLXS</td>\n","      <td>USD</td>\n","      <td>A</td>\n","      <td>334412</td>\n","    </tr>\n","    <tr>\n","      <th>2</th>\n","      <td>HNGR</td>\n","      <td>-0.435712</td>\n","      <td>0.411425</td>\n","      <td>16456</td>\n","      <td>20191231</td>\n","      <td>2019.0</td>\n","      <td>INDL</td>\n","      <td>C</td>\n","      <td>D</td>\n","      <td>STD</td>\n","      <td>HNGR</td>\n","      <td>USD</td>\n","      <td>A</td>\n","      <td>3510</td>\n","      <td>16456</td>\n","      <td>20191231</td>\n","      <td>2019.0</td>\n","      <td>INDL</td>\n","      <td>C</td>\n","      <td>D</td>\n","      <td>STD</td>\n","      <td>HNGR</td>\n","      <td>USD</td>\n","      <td>A</td>\n","      <td>621498</td>\n","    </tr>\n","    <tr>\n","      <th>3</th>\n","      <td>ORCL</td>\n","      <td>-0.083596</td>\n","      <td>0.356166</td>\n","      <td>12142</td>\n","      <td>20200531</td>\n","      <td>2019.0</td>\n","      <td>INDL</td>\n","      <td>C</td>\n","      <td>D</td>\n","      <td>STD</td>\n","      <td>ORCL</td>\n","      <td>USD</td>\n","      <td>A</td>\n","      <td>4510</td>\n","      <td>12142</td>\n","      <td>20200531</td>\n","      <td>2019.0</td>\n","      <td>INDL</td>\n","      <td>C</td>\n","      <td>D</td>\n","      <td>STD</td>\n","      <td>ORCL</td>\n","      <td>USD</td>\n","      <td>A</td>\n","      <td>519130</td>\n","    </tr>\n","    <tr>\n","      <th>4</th>\n","      <td>MSFT</td>\n","      <td>0.003212</td>\n","      <td>0.421139</td>\n","      <td>12141</td>\n","      <td>20190630</td>\n","      <td>2019.0</td>\n","      <td>INDL</td>\n","      <td>C</td>\n","      <td>D</td>\n","      <td>STD</td>\n","      <td>MSFT</td>\n","      <td>USD</td>\n","      <td>A</td>\n","      <td>4510</td>\n","      <td>12141</td>\n","      <td>20190630</td>\n","      <td>2019.0</td>\n","      <td>INDL</td>\n","      <td>C</td>\n","      <td>D</td>\n","      <td>STD</td>\n","      <td>MSFT</td>\n","      <td>USD</td>\n","      <td>A</td>\n","      <td>511210</td>\n","    </tr>\n","  </tbody>\n","</table>\n","</div>"],"text/plain":["  TICKER  RetEarly2020  RetLate2020  gvkey_x  ...  tic_y  curcd_y costat_y   naics\n","0   JJSF     -0.340234     0.300340    12825  ...   JJSF      USD        A  311812\n","1   PLXS     -0.290876     0.433469    12945  ...   PLXS      USD        A  334412\n","2   HNGR     -0.435712     0.411425    16456  ...   HNGR      USD        A  621498\n","3   ORCL     -0.083596     0.356166    12142  ...   ORCL      USD        A  519130\n","4   MSFT      0.003212     0.421139    12141  ...   MSFT      USD        A  511210\n","\n","[5 rows x 25 columns]"]},"metadata":{"tags":[]},"execution_count":226}]},{"cell_type":"code","metadata":{"colab":{"base_uri":"https://localhost:8080/"},"id":"1cDZBMgXkZB_","executionInfo":{"status":"ok","timestamp":1625713111880,"user_tz":240,"elapsed":13,"user":{"displayName":"Tzu-Hua Huang","photoUrl":"https://lh3.googleusercontent.com/a-/AOh14GgbXOdZWwXGVTDuMHbeSme5C5oKAeF6I6ojY6qc=s64","userId":"01458545184256376706"}},"outputId":"db9f54df-5ff8-41fb-e71e-a10e038b08fb"},"source":["df.columns"],"execution_count":227,"outputs":[{"output_type":"execute_result","data":{"text/plain":["Index(['TICKER', 'RetEarly2020', 'RetLate2020', 'gvkey_x', 'datadate_x',\n","       'fyear_x', 'indfmt_x', 'consol_x', 'popsrc_x', 'datafmt_x', 'tic_x',\n","       'curcd_x', 'costat_x', 'ggroup', 'gvkey_y', 'datadate_y', 'fyear_y',\n","       'indfmt_y', 'consol_y', 'popsrc_y', 'datafmt_y', 'tic_y', 'curcd_y',\n","       'costat_y', 'naics'],\n","      dtype='object')"]},"metadata":{"tags":[]},"execution_count":227}]},{"cell_type":"code","metadata":{"colab":{"base_uri":"https://localhost:8080/","height":204},"id":"2uPdJ8opkQbF","executionInfo":{"status":"ok","timestamp":1625713111880,"user_tz":240,"elapsed":12,"user":{"displayName":"Tzu-Hua Huang","photoUrl":"https://lh3.googleusercontent.com/a-/AOh14GgbXOdZWwXGVTDuMHbeSme5C5oKAeF6I6ojY6qc=s64","userId":"01458545184256376706"}},"outputId":"4a45c007-1319-4147-ded9-fbfecebd0d1f"},"source":["# extract required columns\n","df = df.loc[:, ['TICKER', 'RetEarly2020', 'RetLate2020', 'ggroup', 'naics']]\n","df.head()"],"execution_count":228,"outputs":[{"output_type":"execute_result","data":{"text/html":["<div>\n","<style scoped>\n","    .dataframe tbody tr th:only-of-type {\n","        vertical-align: middle;\n","    }\n","\n","    .dataframe tbody tr th {\n","        vertical-align: top;\n","    }\n","\n","    .dataframe thead th {\n","        text-align: right;\n","    }\n","</style>\n","<table border=\"1\" class=\"dataframe\">\n","  <thead>\n","    <tr style=\"text-align: right;\">\n","      <th></th>\n","      <th>TICKER</th>\n","      <th>RetEarly2020</th>\n","      <th>RetLate2020</th>\n","      <th>ggroup</th>\n","      <th>naics</th>\n","    </tr>\n","  </thead>\n","  <tbody>\n","    <tr>\n","      <th>0</th>\n","      <td>JJSF</td>\n","      <td>-0.340234</td>\n","      <td>0.300340</td>\n","      <td>3020</td>\n","      <td>311812</td>\n","    </tr>\n","    <tr>\n","      <th>1</th>\n","      <td>PLXS</td>\n","      <td>-0.290876</td>\n","      <td>0.433469</td>\n","      <td>4520</td>\n","      <td>334412</td>\n","    </tr>\n","    <tr>\n","      <th>2</th>\n","      <td>HNGR</td>\n","      <td>-0.435712</td>\n","      <td>0.411425</td>\n","      <td>3510</td>\n","      <td>621498</td>\n","    </tr>\n","    <tr>\n","      <th>3</th>\n","      <td>ORCL</td>\n","      <td>-0.083596</td>\n","      <td>0.356166</td>\n","      <td>4510</td>\n","      <td>519130</td>\n","    </tr>\n","    <tr>\n","      <th>4</th>\n","      <td>MSFT</td>\n","      <td>0.003212</td>\n","      <td>0.421139</td>\n","      <td>4510</td>\n","      <td>511210</td>\n","    </tr>\n","  </tbody>\n","</table>\n","</div>"],"text/plain":["  TICKER  RetEarly2020  RetLate2020  ggroup   naics\n","0   JJSF     -0.340234     0.300340    3020  311812\n","1   PLXS     -0.290876     0.433469    4520  334412\n","2   HNGR     -0.435712     0.411425    3510  621498\n","3   ORCL     -0.083596     0.356166    4510  519130\n","4   MSFT      0.003212     0.421139    4510  511210"]},"metadata":{"tags":[]},"execution_count":228}]},{"cell_type":"code","metadata":{"colab":{"base_uri":"https://localhost:8080/"},"id":"Fhr5E7BQfZ-Q","executionInfo":{"status":"ok","timestamp":1625713112160,"user_tz":240,"elapsed":290,"user":{"displayName":"Tzu-Hua Huang","photoUrl":"https://lh3.googleusercontent.com/a-/AOh14GgbXOdZWwXGVTDuMHbeSme5C5oKAeF6I6ojY6qc=s64","userId":"01458545184256376706"}},"outputId":"a4382845-1c77-417e-f2ff-a5fa1f19ee8e"},"source":["# save data to csv file\n","df.to_csv('Return2020_w_Industry_Codes.csv', index=False)\n","!cp Return2020_w_Industry_Codes.csv \"/content/gdrive/My Drive/Colab Notebooks/BA870/Individual Project\"\n","!ls"],"execution_count":229,"outputs":[{"output_type":"stream","text":["cp: 'Return2020_w_Industry_Codes.csv' and '/content/gdrive/My Drive/Colab Notebooks/BA870/Individual Project/Return2020_w_Industry_Codes.csv' are the same file\n","'Individual Project- Tzu-Hua Huang.ipynb'   tempGICS.csv\n"," new_file_in_working_directory.txt\t    tempNAICS.csv\n"," Return2020.csv\t\t\t\t    tempSIC.csv\n"," Return2020_w_Industry_Codes.csv\t    WRDS_Variables.csv\n"],"name":"stdout"}]},{"cell_type":"markdown","metadata":{"id":"M8s1SWxG_xKN"},"source":["## Step 2. Create industry indicators\n"]},{"cell_type":"markdown","metadata":{"id":"hnEZ3OiYlnST"},"source":["### 2.a GICS GGROUP\n","Create Industry Indicator (category) variables for the 24 Industry categories based on GICS GGROUP."]},{"cell_type":"code","metadata":{"colab":{"base_uri":"https://localhost:8080/","height":204},"id":"8L5836KX_935","executionInfo":{"status":"ok","timestamp":1625713112161,"user_tz":240,"elapsed":28,"user":{"displayName":"Tzu-Hua Huang","photoUrl":"https://lh3.googleusercontent.com/a-/AOh14GgbXOdZWwXGVTDuMHbeSme5C5oKAeF6I6ojY6qc=s64","userId":"01458545184256376706"}},"outputId":"244632a0-f8f3-4dea-a1ab-a01f56b963f5"},"source":["# extract necessary columns\n","df_gics = pd.read_csv('Return2020_w_Industry_Codes.csv')\n","df_gics = df_gics.loc[:, ['TICKER', 'RetEarly2020', 'RetLate2020', 'ggroup']]\n","df_gics.head()"],"execution_count":230,"outputs":[{"output_type":"execute_result","data":{"text/html":["<div>\n","<style scoped>\n","    .dataframe tbody tr th:only-of-type {\n","        vertical-align: middle;\n","    }\n","\n","    .dataframe tbody tr th {\n","        vertical-align: top;\n","    }\n","\n","    .dataframe thead th {\n","        text-align: right;\n","    }\n","</style>\n","<table border=\"1\" class=\"dataframe\">\n","  <thead>\n","    <tr style=\"text-align: right;\">\n","      <th></th>\n","      <th>TICKER</th>\n","      <th>RetEarly2020</th>\n","      <th>RetLate2020</th>\n","      <th>ggroup</th>\n","    </tr>\n","  </thead>\n","  <tbody>\n","    <tr>\n","      <th>0</th>\n","      <td>JJSF</td>\n","      <td>-0.340234</td>\n","      <td>0.300340</td>\n","      <td>3020</td>\n","    </tr>\n","    <tr>\n","      <th>1</th>\n","      <td>PLXS</td>\n","      <td>-0.290876</td>\n","      <td>0.433469</td>\n","      <td>4520</td>\n","    </tr>\n","    <tr>\n","      <th>2</th>\n","      <td>HNGR</td>\n","      <td>-0.435712</td>\n","      <td>0.411425</td>\n","      <td>3510</td>\n","    </tr>\n","    <tr>\n","      <th>3</th>\n","      <td>ORCL</td>\n","      <td>-0.083596</td>\n","      <td>0.356166</td>\n","      <td>4510</td>\n","    </tr>\n","    <tr>\n","      <th>4</th>\n","      <td>MSFT</td>\n","      <td>0.003212</td>\n","      <td>0.421139</td>\n","      <td>4510</td>\n","    </tr>\n","  </tbody>\n","</table>\n","</div>"],"text/plain":["  TICKER  RetEarly2020  RetLate2020  ggroup\n","0   JJSF     -0.340234     0.300340    3020\n","1   PLXS     -0.290876     0.433469    4520\n","2   HNGR     -0.435712     0.411425    3510\n","3   ORCL     -0.083596     0.356166    4510\n","4   MSFT      0.003212     0.421139    4510"]},"metadata":{"tags":[]},"execution_count":230}]},{"cell_type":"code","metadata":{"colab":{"base_uri":"https://localhost:8080/"},"id":"HmTeJ3immeoB","executionInfo":{"status":"ok","timestamp":1625713112171,"user_tz":240,"elapsed":28,"user":{"displayName":"Tzu-Hua Huang","photoUrl":"https://lh3.googleusercontent.com/a-/AOh14GgbXOdZWwXGVTDuMHbeSme5C5oKAeF6I6ojY6qc=s64","userId":"01458545184256376706"}},"outputId":"8d83dbb9-ac9f-4ba0-9d13-4e14c9ee1c56"},"source":["df_gics.info()"],"execution_count":231,"outputs":[{"output_type":"stream","text":["<class 'pandas.core.frame.DataFrame'>\n","RangeIndex: 2665 entries, 0 to 2664\n","Data columns (total 4 columns):\n"," #   Column        Non-Null Count  Dtype  \n","---  ------        --------------  -----  \n"," 0   TICKER        2665 non-null   object \n"," 1   RetEarly2020  2665 non-null   float64\n"," 2   RetLate2020   2665 non-null   float64\n"," 3   ggroup        2665 non-null   int64  \n","dtypes: float64(2), int64(1), object(1)\n","memory usage: 83.4+ KB\n"],"name":"stdout"}]},{"cell_type":"code","metadata":{"colab":{"base_uri":"https://localhost:8080/","height":224},"id":"6_8oVcvWCQPd","executionInfo":{"status":"ok","timestamp":1625713112288,"user_tz":240,"elapsed":135,"user":{"displayName":"Tzu-Hua Huang","photoUrl":"https://lh3.googleusercontent.com/a-/AOh14GgbXOdZWwXGVTDuMHbeSme5C5oKAeF6I6ojY6qc=s64","userId":"01458545184256376706"}},"outputId":"43dc820d-505f-4189-e8a7-6b39e3146a23"},"source":["# create industry indicators\n","df_gics = pd.get_dummies(df_gics, columns=['ggroup'], dtype=int)\n","df_gics.head()"],"execution_count":232,"outputs":[{"output_type":"execute_result","data":{"text/html":["<div>\n","<style scoped>\n","    .dataframe tbody tr th:only-of-type {\n","        vertical-align: middle;\n","    }\n","\n","    .dataframe tbody tr th {\n","        vertical-align: top;\n","    }\n","\n","    .dataframe thead th {\n","        text-align: right;\n","    }\n","</style>\n","<table border=\"1\" class=\"dataframe\">\n","  <thead>\n","    <tr style=\"text-align: right;\">\n","      <th></th>\n","      <th>TICKER</th>\n","      <th>RetEarly2020</th>\n","      <th>RetLate2020</th>\n","      <th>ggroup_1010</th>\n","      <th>ggroup_1510</th>\n","      <th>ggroup_2010</th>\n","      <th>ggroup_2020</th>\n","      <th>ggroup_2030</th>\n","      <th>ggroup_2510</th>\n","      <th>ggroup_2520</th>\n","      <th>ggroup_2530</th>\n","      <th>ggroup_2550</th>\n","      <th>ggroup_3010</th>\n","      <th>ggroup_3020</th>\n","      <th>ggroup_3030</th>\n","      <th>ggroup_3510</th>\n","      <th>ggroup_3520</th>\n","      <th>ggroup_4010</th>\n","      <th>ggroup_4020</th>\n","      <th>ggroup_4030</th>\n","      <th>ggroup_4510</th>\n","      <th>ggroup_4520</th>\n","      <th>ggroup_4530</th>\n","      <th>ggroup_5010</th>\n","      <th>ggroup_5020</th>\n","      <th>ggroup_5510</th>\n","      <th>ggroup_6010</th>\n","    </tr>\n","  </thead>\n","  <tbody>\n","    <tr>\n","      <th>0</th>\n","      <td>JJSF</td>\n","      <td>-0.340234</td>\n","      <td>0.300340</td>\n","      <td>0</td>\n","      <td>0</td>\n","      <td>0</td>\n","      <td>0</td>\n","      <td>0</td>\n","      <td>0</td>\n","      <td>0</td>\n","      <td>0</td>\n","      <td>0</td>\n","      <td>0</td>\n","      <td>1</td>\n","      <td>0</td>\n","      <td>0</td>\n","      <td>0</td>\n","      <td>0</td>\n","      <td>0</td>\n","      <td>0</td>\n","      <td>0</td>\n","      <td>0</td>\n","      <td>0</td>\n","      <td>0</td>\n","      <td>0</td>\n","      <td>0</td>\n","      <td>0</td>\n","    </tr>\n","    <tr>\n","      <th>1</th>\n","      <td>PLXS</td>\n","      <td>-0.290876</td>\n","      <td>0.433469</td>\n","      <td>0</td>\n","      <td>0</td>\n","      <td>0</td>\n","      <td>0</td>\n","      <td>0</td>\n","      <td>0</td>\n","      <td>0</td>\n","      <td>0</td>\n","      <td>0</td>\n","      <td>0</td>\n","      <td>0</td>\n","      <td>0</td>\n","      <td>0</td>\n","      <td>0</td>\n","      <td>0</td>\n","      <td>0</td>\n","      <td>0</td>\n","      <td>0</td>\n","      <td>1</td>\n","      <td>0</td>\n","      <td>0</td>\n","      <td>0</td>\n","      <td>0</td>\n","      <td>0</td>\n","    </tr>\n","    <tr>\n","      <th>2</th>\n","      <td>HNGR</td>\n","      <td>-0.435712</td>\n","      <td>0.411425</td>\n","      <td>0</td>\n","      <td>0</td>\n","      <td>0</td>\n","      <td>0</td>\n","      <td>0</td>\n","      <td>0</td>\n","      <td>0</td>\n","      <td>0</td>\n","      <td>0</td>\n","      <td>0</td>\n","      <td>0</td>\n","      <td>0</td>\n","      <td>1</td>\n","      <td>0</td>\n","      <td>0</td>\n","      <td>0</td>\n","      <td>0</td>\n","      <td>0</td>\n","      <td>0</td>\n","      <td>0</td>\n","      <td>0</td>\n","      <td>0</td>\n","      <td>0</td>\n","      <td>0</td>\n","    </tr>\n","    <tr>\n","      <th>3</th>\n","      <td>ORCL</td>\n","      <td>-0.083596</td>\n","      <td>0.356166</td>\n","      <td>0</td>\n","      <td>0</td>\n","      <td>0</td>\n","      <td>0</td>\n","      <td>0</td>\n","      <td>0</td>\n","      <td>0</td>\n","      <td>0</td>\n","      <td>0</td>\n","      <td>0</td>\n","      <td>0</td>\n","      <td>0</td>\n","      <td>0</td>\n","      <td>0</td>\n","      <td>0</td>\n","      <td>0</td>\n","      <td>0</td>\n","      <td>1</td>\n","      <td>0</td>\n","      <td>0</td>\n","      <td>0</td>\n","      <td>0</td>\n","      <td>0</td>\n","      <td>0</td>\n","    </tr>\n","    <tr>\n","      <th>4</th>\n","      <td>MSFT</td>\n","      <td>0.003212</td>\n","      <td>0.421139</td>\n","      <td>0</td>\n","      <td>0</td>\n","      <td>0</td>\n","      <td>0</td>\n","      <td>0</td>\n","      <td>0</td>\n","      <td>0</td>\n","      <td>0</td>\n","      <td>0</td>\n","      <td>0</td>\n","      <td>0</td>\n","      <td>0</td>\n","      <td>0</td>\n","      <td>0</td>\n","      <td>0</td>\n","      <td>0</td>\n","      <td>0</td>\n","      <td>1</td>\n","      <td>0</td>\n","      <td>0</td>\n","      <td>0</td>\n","      <td>0</td>\n","      <td>0</td>\n","      <td>0</td>\n","    </tr>\n","  </tbody>\n","</table>\n","</div>"],"text/plain":["  TICKER  RetEarly2020  RetLate2020  ...  ggroup_5020  ggroup_5510  ggroup_6010\n","0   JJSF     -0.340234     0.300340  ...            0            0            0\n","1   PLXS     -0.290876     0.433469  ...            0            0            0\n","2   HNGR     -0.435712     0.411425  ...            0            0            0\n","3   ORCL     -0.083596     0.356166  ...            0            0            0\n","4   MSFT      0.003212     0.421139  ...            0            0            0\n","\n","[5 rows x 27 columns]"]},"metadata":{"tags":[]},"execution_count":232}]},{"cell_type":"code","metadata":{"colab":{"base_uri":"https://localhost:8080/"},"id":"ubPKN_asCw3m","executionInfo":{"status":"ok","timestamp":1625713112289,"user_tz":240,"elapsed":30,"user":{"displayName":"Tzu-Hua Huang","photoUrl":"https://lh3.googleusercontent.com/a-/AOh14GgbXOdZWwXGVTDuMHbeSme5C5oKAeF6I6ojY6qc=s64","userId":"01458545184256376706"}},"outputId":"42736302-c1ab-4e55-aa1a-7729832f54fa"},"source":["df_gics.info()"],"execution_count":233,"outputs":[{"output_type":"stream","text":["<class 'pandas.core.frame.DataFrame'>\n","RangeIndex: 2665 entries, 0 to 2664\n","Data columns (total 27 columns):\n"," #   Column        Non-Null Count  Dtype  \n","---  ------        --------------  -----  \n"," 0   TICKER        2665 non-null   object \n"," 1   RetEarly2020  2665 non-null   float64\n"," 2   RetLate2020   2665 non-null   float64\n"," 3   ggroup_1010   2665 non-null   int64  \n"," 4   ggroup_1510   2665 non-null   int64  \n"," 5   ggroup_2010   2665 non-null   int64  \n"," 6   ggroup_2020   2665 non-null   int64  \n"," 7   ggroup_2030   2665 non-null   int64  \n"," 8   ggroup_2510   2665 non-null   int64  \n"," 9   ggroup_2520   2665 non-null   int64  \n"," 10  ggroup_2530   2665 non-null   int64  \n"," 11  ggroup_2550   2665 non-null   int64  \n"," 12  ggroup_3010   2665 non-null   int64  \n"," 13  ggroup_3020   2665 non-null   int64  \n"," 14  ggroup_3030   2665 non-null   int64  \n"," 15  ggroup_3510   2665 non-null   int64  \n"," 16  ggroup_3520   2665 non-null   int64  \n"," 17  ggroup_4010   2665 non-null   int64  \n"," 18  ggroup_4020   2665 non-null   int64  \n"," 19  ggroup_4030   2665 non-null   int64  \n"," 20  ggroup_4510   2665 non-null   int64  \n"," 21  ggroup_4520   2665 non-null   int64  \n"," 22  ggroup_4530   2665 non-null   int64  \n"," 23  ggroup_5010   2665 non-null   int64  \n"," 24  ggroup_5020   2665 non-null   int64  \n"," 25  ggroup_5510   2665 non-null   int64  \n"," 26  ggroup_6010   2665 non-null   int64  \n","dtypes: float64(2), int64(24), object(1)\n","memory usage: 562.3+ KB\n"],"name":"stdout"}]},{"cell_type":"markdown","metadata":{"id":"qNtaImrkl5xT"},"source":["### 2.b NAICS\n","Create Industry Indicator (category) variables for the industry categories based on NAICS."]},{"cell_type":"code","metadata":{"colab":{"base_uri":"https://localhost:8080/","height":204},"id":"gVMRTPpol5xU","executionInfo":{"status":"ok","timestamp":1625713112290,"user_tz":240,"elapsed":21,"user":{"displayName":"Tzu-Hua Huang","photoUrl":"https://lh3.googleusercontent.com/a-/AOh14GgbXOdZWwXGVTDuMHbeSme5C5oKAeF6I6ojY6qc=s64","userId":"01458545184256376706"}},"outputId":"a95f45b0-af83-4e12-81e7-895f849a0f10"},"source":["# extract necessary columns\n","df_naics = pd.read_csv('Return2020_w_Industry_Codes.csv')\n","df_naics = df_naics.loc[:, ['TICKER', 'RetEarly2020', 'RetLate2020', 'naics']]\n","df_naics.head()"],"execution_count":234,"outputs":[{"output_type":"execute_result","data":{"text/html":["<div>\n","<style scoped>\n","    .dataframe tbody tr th:only-of-type {\n","        vertical-align: middle;\n","    }\n","\n","    .dataframe tbody tr th {\n","        vertical-align: top;\n","    }\n","\n","    .dataframe thead th {\n","        text-align: right;\n","    }\n","</style>\n","<table border=\"1\" class=\"dataframe\">\n","  <thead>\n","    <tr style=\"text-align: right;\">\n","      <th></th>\n","      <th>TICKER</th>\n","      <th>RetEarly2020</th>\n","      <th>RetLate2020</th>\n","      <th>naics</th>\n","    </tr>\n","  </thead>\n","  <tbody>\n","    <tr>\n","      <th>0</th>\n","      <td>JJSF</td>\n","      <td>-0.340234</td>\n","      <td>0.300340</td>\n","      <td>311812</td>\n","    </tr>\n","    <tr>\n","      <th>1</th>\n","      <td>PLXS</td>\n","      <td>-0.290876</td>\n","      <td>0.433469</td>\n","      <td>334412</td>\n","    </tr>\n","    <tr>\n","      <th>2</th>\n","      <td>HNGR</td>\n","      <td>-0.435712</td>\n","      <td>0.411425</td>\n","      <td>621498</td>\n","    </tr>\n","    <tr>\n","      <th>3</th>\n","      <td>ORCL</td>\n","      <td>-0.083596</td>\n","      <td>0.356166</td>\n","      <td>519130</td>\n","    </tr>\n","    <tr>\n","      <th>4</th>\n","      <td>MSFT</td>\n","      <td>0.003212</td>\n","      <td>0.421139</td>\n","      <td>511210</td>\n","    </tr>\n","  </tbody>\n","</table>\n","</div>"],"text/plain":["  TICKER  RetEarly2020  RetLate2020   naics\n","0   JJSF     -0.340234     0.300340  311812\n","1   PLXS     -0.290876     0.433469  334412\n","2   HNGR     -0.435712     0.411425  621498\n","3   ORCL     -0.083596     0.356166  519130\n","4   MSFT      0.003212     0.421139  511210"]},"metadata":{"tags":[]},"execution_count":234}]},{"cell_type":"code","metadata":{"colab":{"base_uri":"https://localhost:8080/"},"id":"UTGBnWQKmSQG","executionInfo":{"status":"ok","timestamp":1625713112441,"user_tz":240,"elapsed":170,"user":{"displayName":"Tzu-Hua Huang","photoUrl":"https://lh3.googleusercontent.com/a-/AOh14GgbXOdZWwXGVTDuMHbeSme5C5oKAeF6I6ojY6qc=s64","userId":"01458545184256376706"}},"outputId":"63a34071-285d-44c0-ca2c-e4177a0ce524"},"source":["df_naics.info()"],"execution_count":235,"outputs":[{"output_type":"stream","text":["<class 'pandas.core.frame.DataFrame'>\n","RangeIndex: 2665 entries, 0 to 2664\n","Data columns (total 4 columns):\n"," #   Column        Non-Null Count  Dtype  \n","---  ------        --------------  -----  \n"," 0   TICKER        2665 non-null   object \n"," 1   RetEarly2020  2665 non-null   float64\n"," 2   RetLate2020   2665 non-null   float64\n"," 3   naics         2665 non-null   int64  \n","dtypes: float64(2), int64(1), object(1)\n","memory usage: 83.4+ KB\n"],"name":"stdout"}]},{"cell_type":"code","metadata":{"colab":{"base_uri":"https://localhost:8080/"},"id":"ucEq42dtBCQx","executionInfo":{"status":"ok","timestamp":1625713112442,"user_tz":240,"elapsed":15,"user":{"displayName":"Tzu-Hua Huang","photoUrl":"https://lh3.googleusercontent.com/a-/AOh14GgbXOdZWwXGVTDuMHbeSme5C5oKAeF6I6ojY6qc=s64","userId":"01458545184256376706"}},"outputId":"52247e03-7869-4a8e-efb1-0fc91888b83a"},"source":["df_naics[['naics']] = df_naics[['naics']].astype(str)\n","df_naics['Sector_naics'] = df_naics['naics'].str[:2]\n","df_naics.info()"],"execution_count":236,"outputs":[{"output_type":"stream","text":["<class 'pandas.core.frame.DataFrame'>\n","RangeIndex: 2665 entries, 0 to 2664\n","Data columns (total 5 columns):\n"," #   Column        Non-Null Count  Dtype  \n","---  ------        --------------  -----  \n"," 0   TICKER        2665 non-null   object \n"," 1   RetEarly2020  2665 non-null   float64\n"," 2   RetLate2020   2665 non-null   float64\n"," 3   naics         2665 non-null   object \n"," 4   Sector_naics  2665 non-null   object \n","dtypes: float64(2), object(3)\n","memory usage: 104.2+ KB\n"],"name":"stdout"}]},{"cell_type":"code","metadata":{"colab":{"base_uri":"https://localhost:8080/","height":224},"id":"ezIV1cTul5xU","executionInfo":{"status":"ok","timestamp":1625713112443,"user_tz":240,"elapsed":14,"user":{"displayName":"Tzu-Hua Huang","photoUrl":"https://lh3.googleusercontent.com/a-/AOh14GgbXOdZWwXGVTDuMHbeSme5C5oKAeF6I6ojY6qc=s64","userId":"01458545184256376706"}},"outputId":"c37729a6-4eef-4b18-e06d-efd7be517688"},"source":["# create industry indicators\n","df_naics = pd.get_dummies(df_naics, columns=['Sector_naics'], dtype=int)\n","df_naics.head()"],"execution_count":237,"outputs":[{"output_type":"execute_result","data":{"text/html":["<div>\n","<style scoped>\n","    .dataframe tbody tr th:only-of-type {\n","        vertical-align: middle;\n","    }\n","\n","    .dataframe tbody tr th {\n","        vertical-align: top;\n","    }\n","\n","    .dataframe thead th {\n","        text-align: right;\n","    }\n","</style>\n","<table border=\"1\" class=\"dataframe\">\n","  <thead>\n","    <tr style=\"text-align: right;\">\n","      <th></th>\n","      <th>TICKER</th>\n","      <th>RetEarly2020</th>\n","      <th>RetLate2020</th>\n","      <th>naics</th>\n","      <th>Sector_naics_11</th>\n","      <th>Sector_naics_21</th>\n","      <th>Sector_naics_22</th>\n","      <th>Sector_naics_23</th>\n","      <th>Sector_naics_31</th>\n","      <th>Sector_naics_32</th>\n","      <th>Sector_naics_33</th>\n","      <th>Sector_naics_42</th>\n","      <th>Sector_naics_44</th>\n","      <th>Sector_naics_45</th>\n","      <th>Sector_naics_48</th>\n","      <th>Sector_naics_49</th>\n","      <th>Sector_naics_51</th>\n","      <th>Sector_naics_52</th>\n","      <th>Sector_naics_53</th>\n","      <th>Sector_naics_54</th>\n","      <th>Sector_naics_56</th>\n","      <th>Sector_naics_61</th>\n","      <th>Sector_naics_62</th>\n","      <th>Sector_naics_71</th>\n","      <th>Sector_naics_72</th>\n","      <th>Sector_naics_81</th>\n","      <th>Sector_naics_99</th>\n","    </tr>\n","  </thead>\n","  <tbody>\n","    <tr>\n","      <th>0</th>\n","      <td>JJSF</td>\n","      <td>-0.340234</td>\n","      <td>0.300340</td>\n","      <td>311812</td>\n","      <td>0</td>\n","      <td>0</td>\n","      <td>0</td>\n","      <td>0</td>\n","      <td>1</td>\n","      <td>0</td>\n","      <td>0</td>\n","      <td>0</td>\n","      <td>0</td>\n","      <td>0</td>\n","      <td>0</td>\n","      <td>0</td>\n","      <td>0</td>\n","      <td>0</td>\n","      <td>0</td>\n","      <td>0</td>\n","      <td>0</td>\n","      <td>0</td>\n","      <td>0</td>\n","      <td>0</td>\n","      <td>0</td>\n","      <td>0</td>\n","      <td>0</td>\n","    </tr>\n","    <tr>\n","      <th>1</th>\n","      <td>PLXS</td>\n","      <td>-0.290876</td>\n","      <td>0.433469</td>\n","      <td>334412</td>\n","      <td>0</td>\n","      <td>0</td>\n","      <td>0</td>\n","      <td>0</td>\n","      <td>0</td>\n","      <td>0</td>\n","      <td>1</td>\n","      <td>0</td>\n","      <td>0</td>\n","      <td>0</td>\n","      <td>0</td>\n","      <td>0</td>\n","      <td>0</td>\n","      <td>0</td>\n","      <td>0</td>\n","      <td>0</td>\n","      <td>0</td>\n","      <td>0</td>\n","      <td>0</td>\n","      <td>0</td>\n","      <td>0</td>\n","      <td>0</td>\n","      <td>0</td>\n","    </tr>\n","    <tr>\n","      <th>2</th>\n","      <td>HNGR</td>\n","      <td>-0.435712</td>\n","      <td>0.411425</td>\n","      <td>621498</td>\n","      <td>0</td>\n","      <td>0</td>\n","      <td>0</td>\n","      <td>0</td>\n","      <td>0</td>\n","      <td>0</td>\n","      <td>0</td>\n","      <td>0</td>\n","      <td>0</td>\n","      <td>0</td>\n","      <td>0</td>\n","      <td>0</td>\n","      <td>0</td>\n","      <td>0</td>\n","      <td>0</td>\n","      <td>0</td>\n","      <td>0</td>\n","      <td>0</td>\n","      <td>1</td>\n","      <td>0</td>\n","      <td>0</td>\n","      <td>0</td>\n","      <td>0</td>\n","    </tr>\n","    <tr>\n","      <th>3</th>\n","      <td>ORCL</td>\n","      <td>-0.083596</td>\n","      <td>0.356166</td>\n","      <td>519130</td>\n","      <td>0</td>\n","      <td>0</td>\n","      <td>0</td>\n","      <td>0</td>\n","      <td>0</td>\n","      <td>0</td>\n","      <td>0</td>\n","      <td>0</td>\n","      <td>0</td>\n","      <td>0</td>\n","      <td>0</td>\n","      <td>0</td>\n","      <td>1</td>\n","      <td>0</td>\n","      <td>0</td>\n","      <td>0</td>\n","      <td>0</td>\n","      <td>0</td>\n","      <td>0</td>\n","      <td>0</td>\n","      <td>0</td>\n","      <td>0</td>\n","      <td>0</td>\n","    </tr>\n","    <tr>\n","      <th>4</th>\n","      <td>MSFT</td>\n","      <td>0.003212</td>\n","      <td>0.421139</td>\n","      <td>511210</td>\n","      <td>0</td>\n","      <td>0</td>\n","      <td>0</td>\n","      <td>0</td>\n","      <td>0</td>\n","      <td>0</td>\n","      <td>0</td>\n","      <td>0</td>\n","      <td>0</td>\n","      <td>0</td>\n","      <td>0</td>\n","      <td>0</td>\n","      <td>1</td>\n","      <td>0</td>\n","      <td>0</td>\n","      <td>0</td>\n","      <td>0</td>\n","      <td>0</td>\n","      <td>0</td>\n","      <td>0</td>\n","      <td>0</td>\n","      <td>0</td>\n","      <td>0</td>\n","    </tr>\n","  </tbody>\n","</table>\n","</div>"],"text/plain":["  TICKER  RetEarly2020  ...  Sector_naics_81 Sector_naics_99\n","0   JJSF     -0.340234  ...                0               0\n","1   PLXS     -0.290876  ...                0               0\n","2   HNGR     -0.435712  ...                0               0\n","3   ORCL     -0.083596  ...                0               0\n","4   MSFT      0.003212  ...                0               0\n","\n","[5 rows x 27 columns]"]},"metadata":{"tags":[]},"execution_count":237}]},{"cell_type":"code","metadata":{"colab":{"base_uri":"https://localhost:8080/"},"id":"wzUUwhzhl5xU","executionInfo":{"status":"ok","timestamp":1625713112443,"user_tz":240,"elapsed":13,"user":{"displayName":"Tzu-Hua Huang","photoUrl":"https://lh3.googleusercontent.com/a-/AOh14GgbXOdZWwXGVTDuMHbeSme5C5oKAeF6I6ojY6qc=s64","userId":"01458545184256376706"}},"outputId":"a10dc8fd-268a-44c2-9eb2-bfd4c3791677"},"source":["df_naics.info()"],"execution_count":238,"outputs":[{"output_type":"stream","text":["<class 'pandas.core.frame.DataFrame'>\n","RangeIndex: 2665 entries, 0 to 2664\n","Data columns (total 27 columns):\n"," #   Column           Non-Null Count  Dtype  \n","---  ------           --------------  -----  \n"," 0   TICKER           2665 non-null   object \n"," 1   RetEarly2020     2665 non-null   float64\n"," 2   RetLate2020      2665 non-null   float64\n"," 3   naics            2665 non-null   object \n"," 4   Sector_naics_11  2665 non-null   int64  \n"," 5   Sector_naics_21  2665 non-null   int64  \n"," 6   Sector_naics_22  2665 non-null   int64  \n"," 7   Sector_naics_23  2665 non-null   int64  \n"," 8   Sector_naics_31  2665 non-null   int64  \n"," 9   Sector_naics_32  2665 non-null   int64  \n"," 10  Sector_naics_33  2665 non-null   int64  \n"," 11  Sector_naics_42  2665 non-null   int64  \n"," 12  Sector_naics_44  2665 non-null   int64  \n"," 13  Sector_naics_45  2665 non-null   int64  \n"," 14  Sector_naics_48  2665 non-null   int64  \n"," 15  Sector_naics_49  2665 non-null   int64  \n"," 16  Sector_naics_51  2665 non-null   int64  \n"," 17  Sector_naics_52  2665 non-null   int64  \n"," 18  Sector_naics_53  2665 non-null   int64  \n"," 19  Sector_naics_54  2665 non-null   int64  \n"," 20  Sector_naics_56  2665 non-null   int64  \n"," 21  Sector_naics_61  2665 non-null   int64  \n"," 22  Sector_naics_62  2665 non-null   int64  \n"," 23  Sector_naics_71  2665 non-null   int64  \n"," 24  Sector_naics_72  2665 non-null   int64  \n"," 25  Sector_naics_81  2665 non-null   int64  \n"," 26  Sector_naics_99  2665 non-null   int64  \n","dtypes: float64(2), int64(23), object(2)\n","memory usage: 562.3+ KB\n"],"name":"stdout"}]},{"cell_type":"markdown","metadata":{"id":"mNLEdMTa_-sv"},"source":["## Step 3. Calculation and estimation\n","For both *ReturnEarly2020* and *ReturnLate2020*, calculate and estimate the following:"]},{"cell_type":"markdown","metadata":{"id":"cV5yEto3oopn"},"source":["### 3.a GICS GGROUP"]},{"cell_type":"markdown","metadata":{"id":"lejv-nj7UAhj"},"source":["#### i. Average stock returns\n","What are the average stock returns for each industry classification (24 industry groups)?\n","\n","1. Do the highest and lowest return industries make economic sense for early and late 2020? Explain in detail."]},{"cell_type":"code","metadata":{"colab":{"base_uri":"https://localhost:8080/","height":824},"id":"MrD03McOUvTx","executionInfo":{"status":"ok","timestamp":1625713112576,"user_tz":240,"elapsed":140,"user":{"displayName":"Tzu-Hua Huang","photoUrl":"https://lh3.googleusercontent.com/a-/AOh14GgbXOdZWwXGVTDuMHbeSme5C5oKAeF6I6ojY6qc=s64","userId":"01458545184256376706"}},"outputId":"01882253-bad1-413f-e8e5-dda43756a662"},"source":["df_gics_AvgRet = df_gics.drop(columns=['TICKER'])\n","df_gics_AvgRet = df_gics_AvgRet.melt(id_vars=['RetEarly2020', 'RetLate2020']).groupby(['variable', 'value']).mean()\n","df_gics_AvgRet.query('value == 1')"],"execution_count":239,"outputs":[{"output_type":"execute_result","data":{"text/html":["<div>\n","<style scoped>\n","    .dataframe tbody tr th:only-of-type {\n","        vertical-align: middle;\n","    }\n","\n","    .dataframe tbody tr th {\n","        vertical-align: top;\n","    }\n","\n","    .dataframe thead th {\n","        text-align: right;\n","    }\n","</style>\n","<table border=\"1\" class=\"dataframe\">\n","  <thead>\n","    <tr style=\"text-align: right;\">\n","      <th></th>\n","      <th></th>\n","      <th>RetEarly2020</th>\n","      <th>RetLate2020</th>\n","    </tr>\n","    <tr>\n","      <th>variable</th>\n","      <th>value</th>\n","      <th></th>\n","      <th></th>\n","    </tr>\n","  </thead>\n","  <tbody>\n","    <tr>\n","      <th>ggroup_1010</th>\n","      <th>1</th>\n","      <td>-0.580745</td>\n","      <td>1.017517</td>\n","    </tr>\n","    <tr>\n","      <th>ggroup_1510</th>\n","      <th>1</th>\n","      <td>-0.350917</td>\n","      <td>0.873224</td>\n","    </tr>\n","    <tr>\n","      <th>ggroup_2010</th>\n","      <th>1</th>\n","      <td>-0.316910</td>\n","      <td>0.898123</td>\n","    </tr>\n","    <tr>\n","      <th>ggroup_2020</th>\n","      <th>1</th>\n","      <td>-0.332514</td>\n","      <td>0.605461</td>\n","    </tr>\n","    <tr>\n","      <th>ggroup_2030</th>\n","      <th>1</th>\n","      <td>-0.326556</td>\n","      <td>0.709745</td>\n","    </tr>\n","    <tr>\n","      <th>ggroup_2510</th>\n","      <th>1</th>\n","      <td>-0.429483</td>\n","      <td>1.811358</td>\n","    </tr>\n","    <tr>\n","      <th>ggroup_2520</th>\n","      <th>1</th>\n","      <td>-0.391515</td>\n","      <td>1.594357</td>\n","    </tr>\n","    <tr>\n","      <th>ggroup_2530</th>\n","      <th>1</th>\n","      <td>-0.448220</td>\n","      <td>1.131115</td>\n","    </tr>\n","    <tr>\n","      <th>ggroup_2550</th>\n","      <th>1</th>\n","      <td>-0.385352</td>\n","      <td>1.780307</td>\n","    </tr>\n","    <tr>\n","      <th>ggroup_3010</th>\n","      <th>1</th>\n","      <td>-0.175851</td>\n","      <td>0.404087</td>\n","    </tr>\n","    <tr>\n","      <th>ggroup_3020</th>\n","      <th>1</th>\n","      <td>-0.181004</td>\n","      <td>0.533474</td>\n","    </tr>\n","    <tr>\n","      <th>ggroup_3030</th>\n","      <th>1</th>\n","      <td>-0.231034</td>\n","      <td>0.611502</td>\n","    </tr>\n","    <tr>\n","      <th>ggroup_3510</th>\n","      <th>1</th>\n","      <td>-0.134281</td>\n","      <td>0.739733</td>\n","    </tr>\n","    <tr>\n","      <th>ggroup_3520</th>\n","      <th>1</th>\n","      <td>-0.141030</td>\n","      <td>0.739747</td>\n","    </tr>\n","    <tr>\n","      <th>ggroup_4010</th>\n","      <th>1</th>\n","      <td>-0.368477</td>\n","      <td>0.433829</td>\n","    </tr>\n","    <tr>\n","      <th>ggroup_4020</th>\n","      <th>1</th>\n","      <td>-0.359551</td>\n","      <td>0.745559</td>\n","    </tr>\n","    <tr>\n","      <th>ggroup_4030</th>\n","      <th>1</th>\n","      <td>-0.249471</td>\n","      <td>0.358440</td>\n","    </tr>\n","    <tr>\n","      <th>ggroup_4510</th>\n","      <th>1</th>\n","      <td>-0.176543</td>\n","      <td>1.035505</td>\n","    </tr>\n","    <tr>\n","      <th>ggroup_4520</th>\n","      <th>1</th>\n","      <td>-0.289550</td>\n","      <td>0.817086</td>\n","    </tr>\n","    <tr>\n","      <th>ggroup_4530</th>\n","      <th>1</th>\n","      <td>-0.226510</td>\n","      <td>1.218551</td>\n","    </tr>\n","    <tr>\n","      <th>ggroup_5010</th>\n","      <th>1</th>\n","      <td>-0.089193</td>\n","      <td>0.490060</td>\n","    </tr>\n","    <tr>\n","      <th>ggroup_5020</th>\n","      <th>1</th>\n","      <td>-0.341396</td>\n","      <td>0.854719</td>\n","    </tr>\n","    <tr>\n","      <th>ggroup_5510</th>\n","      <th>1</th>\n","      <td>-0.147903</td>\n","      <td>0.181212</td>\n","    </tr>\n","    <tr>\n","      <th>ggroup_6010</th>\n","      <th>1</th>\n","      <td>-0.328122</td>\n","      <td>0.456668</td>\n","    </tr>\n","  </tbody>\n","</table>\n","</div>"],"text/plain":["                   RetEarly2020  RetLate2020\n","variable    value                           \n","ggroup_1010 1         -0.580745     1.017517\n","ggroup_1510 1         -0.350917     0.873224\n","ggroup_2010 1         -0.316910     0.898123\n","ggroup_2020 1         -0.332514     0.605461\n","ggroup_2030 1         -0.326556     0.709745\n","ggroup_2510 1         -0.429483     1.811358\n","ggroup_2520 1         -0.391515     1.594357\n","ggroup_2530 1         -0.448220     1.131115\n","ggroup_2550 1         -0.385352     1.780307\n","ggroup_3010 1         -0.175851     0.404087\n","ggroup_3020 1         -0.181004     0.533474\n","ggroup_3030 1         -0.231034     0.611502\n","ggroup_3510 1         -0.134281     0.739733\n","ggroup_3520 1         -0.141030     0.739747\n","ggroup_4010 1         -0.368477     0.433829\n","ggroup_4020 1         -0.359551     0.745559\n","ggroup_4030 1         -0.249471     0.358440\n","ggroup_4510 1         -0.176543     1.035505\n","ggroup_4520 1         -0.289550     0.817086\n","ggroup_4530 1         -0.226510     1.218551\n","ggroup_5010 1         -0.089193     0.490060\n","ggroup_5020 1         -0.341396     0.854719\n","ggroup_5510 1         -0.147903     0.181212\n","ggroup_6010 1         -0.328122     0.456668"]},"metadata":{"tags":[]},"execution_count":239}]},{"cell_type":"code","metadata":{"colab":{"base_uri":"https://localhost:8080/","height":372},"id":"vTSQf-wKVQxr","executionInfo":{"status":"ok","timestamp":1625713113139,"user_tz":240,"elapsed":568,"user":{"displayName":"Tzu-Hua Huang","photoUrl":"https://lh3.googleusercontent.com/a-/AOh14GgbXOdZWwXGVTDuMHbeSme5C5oKAeF6I6ojY6qc=s64","userId":"01458545184256376706"}},"outputId":"dde87928-f8e5-4df4-9a57-533d01729c4c"},"source":["# barplot\n","df_gics_AvgRet.query('value == 1').plot(kind='bar')"],"execution_count":240,"outputs":[{"output_type":"execute_result","data":{"text/plain":["<matplotlib.axes._subplots.AxesSubplot at 0x7f81c22afe50>"]},"metadata":{"tags":[]},"execution_count":240},{"output_type":"display_data","data":{"image/png":"iVBORw0KGgoAAAANSUhEUgAAAXwAAAFSCAYAAADxQWSMAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADh0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uMy4yLjIsIGh0dHA6Ly9tYXRwbG90bGliLm9yZy+WH4yJAAAgAElEQVR4nO3de5wU1Zn/8c8DDEGUi1xUItd4iURRVFAx0aiEn5foGPFGFgMoBiMazSabLPpbozHGH4a4cTWa6KpRicFbImDiBRcV47peQFEQVCSiDiGIGFmJokCe3x9VMzRDd093dU911dT3/XrVa7qr6qlz6nT10zWnq0+ZuyMiIm1fu1pXQERE4qGELyKSEUr4IiIZoYQvIpIRSvgiIhnRodYVKKRXr14+cODAWldDRCRVFixY8J679863LLEJf+DAgcyfP7/W1RARSRUze6vQMnXpiIhkhBK+iEhGKOGLiGREYvvwRSQZNm7cSENDAxs2bKh1VSRHp06d6Nu3L3V1dSXHKOGLSFENDQ106dKFgQMHYma1ro4A7s7atWtpaGhg0KBBJcepS0dEitqwYQM9e/ZUsk8QM6Nnz55l/9elhC8iLVKyT54or4kSvohIRqgPvy25rFuz5+tqUw9p0wZO+WNVt7di6ldbXKd9+/YMGTKETZs2MWjQIKZPn0737t0Lrj9z5kz23HNPvvCFLwAwYcIE5s2bR7duwXukc+fOPP3006XXccUKjj/+eBYvXlxwnTvvvJOrrroKd6dLly788pe/ZL/99gPg4Ycf5sILL2Tz5s2cffbZTJkyBYCxY8cyf/586urqOOigg7jxxhupq6vD3bnwwgt58MEH6dy5M7fddhsHHHBAyfUtRGf4IpJ42223HQsXLmTx4sX06NGD66+/vuj6M2fOZMmSJVvNmzZtGgsXLmThwoVlJftNmzaVtN6gQYOYN28eixYt4pJLLmHSpEkAbN68mfPOO4+HHnqIJUuWMGPGjKa6jR07lldffZVFixbx8ccfc/PNNwPw0EMPsWzZMpYtW8ZNN93EueeeW3J9i1HCF5FUGTFiBCtXrgRg+fLlHHPMMRx44IEcdthhvPrqqzz99NPMnj2b73//+wwdOpTly5cX3NZzzz3HiBEj2H///Tn00EN57bXXALjtttuor6/nqKOOYuTIkVvFHH744SxcuLDp+Ze+9CVeeuklDj30UHbccUcADjnkEBoaGprK2H333fnc5z5Hx44dGTNmDLNmzQLguOOOw8wwMw466KCmmFmzZjFu3DjMjEMOOYQPPviAVatWVdx2SvgSv8u6bT2JlGjz5s3MnTuX+vp6ACZNmsR1113HggUL+NnPfsbkyZM59NBDqa+vbzqj32233QCaPgCGDh3K2LFjAdhrr73405/+xIsvvsjll1/OxRdf3FTWCy+8wH333ce8efO2qsPEiRO57bbbAHj99dfZsGFDU9dNo1tuuYVjjz0WgJUrV9KvX7+mZX379m36wGq0ceNGpk+fzjHHHFNyTBTqwxeRxPv4448ZOnQoK1euZPDgwYwaNYr169fz9NNPc+qppzat98knnxTcxrRp0zjllFO2mrdu3TrGjx/PsmXLMDM2btzYtGzUqFH06NFjm+2ceuqp/PjHP2batGnceuutTJgwYavljz/+OLfccgtPPfVUyfs3efJkDj/8cA477LCSY6JQwheRxGvsw//oo484+uijuf7665kwYQLdu3ffqnulXJdccglHHnkk999/PytWrOCII45oWrb99tvnjencuTOjRo1i1qxZ3HPPPSxYsKBp2csvv8zZZ5/NQw89RM+ePQHYddddeeedd5rWaWhoYNddd216/qMf/Yg1a9Zw4403Ns1rKSYqdemISGp07tyZa6+9lquvvprOnTszaNAg7r33XiD49elLL70EQJcuXfjwww9b3N66deuaEmljN00pzj77bC644AKGDx/e1G//9ttvM3r0aKZPn86ee+7ZtO7w4cNZtmwZb775Jp9++il33XVXU5fUzTffzCOPPMKMGTNo125LOq6vr+eOO+7A3XnmmWfo1q0bffr0Kbl+hegMX0TKUspllK1p//33Z99992XGjBnceeednHvuuVxxxRVs3LiRMWPGsN9++zFmzBi++c1vcu2113LfffcBQR/+FVdc0bSd5557jh/84AeMHz+eK664gq9+tfT9OvDAA+natStnnnlm07zLL7+ctWvXMnnyZAA6dOjA/Pnz6dChA7/4xS84+uij2bx5M2eddRZ77703AN/61rcYMGAAI0aMAGD06NH88Ic/5LjjjuPBBx9k9913p3Pnzvz617+uuN0AzN2rsqFqGzZsmOsGKGVKy3X4aamnALB06VIGDx5c62okyl/+8heOOOIIXn311a3OzOOW77UxswXuPizf+urSEREpwx133MHBBx/MT37yk5om+yjUpSMiUoZx48Yxbty4WlcjknR9PImISGRK+CIiGaGELyKSEUr4IiIZoS9tRaQ81R7/qITLcqsxPPLxxx+/zdAK+XzwwQf89re/bbqevpB33nmHcePGsXr1asyMSZMmceGFFwLw/vvvc/rpp7NixQoGDhzIPffcw4477hhpCOVq0hm+iCReNYZHLtUHH3zADTfc0OJ6HTp04Oqrr2bJkiU888wzXH/99U1lTp06lZEjR7Js2TJGjhzJ1KlTgWhDKFeTEr6IpEq1hkdev349I0eO5IADDmDIkCFNQxZPmTKF5cuXM3ToUL7//e8DwcBrw4cPZ9999+XSSy8FoE+fPk03JenSpQuDBw9uqtesWbMYP348AOPHj2fmzJkAkYZQriZ16Yh++Sqp0Tg88sSJE4FgeORf/epX7LHHHjz77LNMnjyZxx57jPr6+ha7cDp16sT9999P165dee+99zjkkEOor69n6tSpLF68uGlQtjlz5rBs2TKee+453J36+nqefPJJDj/88KZtrVixghdffJGDDz4YgNWrVzeNfbPLLruwevXqbcpvaQjlZ599tsLW2pYSvogkXjWGR27O3bn44ot58sknadeuHStXrsybmOfMmcOcOXPYf//9geA/g2XLljUl/PXr13PyySdzzTXX0LVr123iG29wkivKEMrVoIQvIonXGsMj33nnnaxZs4YFCxZQV1fHwIED2bBhwzbruTsXXXQR55xzzjbLNm7cyMknn8zYsWMZPXp00/ydd96ZVatW0adPH1atWsVOO+3UtCzKEMrVoj58EUmNag6PvG7dOnbaaSfq6up4/PHHeeutt/LGHn300dx6662sX78eCLpf3n33XdydiRMnMnjwYL773e9ute36+npuv/12AG6//XZOPPFEINoQytWkM3wRKU+Nv+OJOjzyOeecw3e+8x0A+vXrxwMPPMAJJ5zAkCFDGDZsGHvttRcAPXv25Itf/CL77LMPxx57LNOmTWPp0qVNQxjvsMMO/OY3v+H1119n+vTpDBkyhKFDhwJw5ZVXctxxxzFlyhROO+00brnlFgYMGMA999wDRBtCuZqqMjyymd0KHA+86+775Fl+BDALeDOc9Xt3v7zYNjU8cgRRv3yN+0tbfUmcKhoeObnKHR65Wmf4twG/AO4oss6f3P34KpUnIiJlqkofvrs/CbxfjW2JiEjriPNL2xFm9pKZPWRmeTunzGySmc03s/lr1qyJsWoiUkxS74yXZVFek7gS/gvAAHffD7gOmJlvJXe/yd2Hufuw3r17x1Q1ESmmU6dOrF27Vkk/QdydtWvX0qlTp7LiYrlKx93/N+fxg2Z2g5n1cvf34ihfRKLr27cvDQ0N6L/uZOnUqRN9+/YtKyaWhG9muwCr3d3N7CCC/yzWxlG2iFSmrq6OQYMG1boaUgVVSfhmNgM4AuhlZg3ApUAdgLv/CjgFONfMNgEfA2Nc/x+KiMSqKgnf3b/ewvJfEFy2KSIiNaKhFUREMkIJX0QkI5TwRUQyQglfRCQjlPBFRDJCCV9EJCOU8EVEMkIJX0QkI5TwRUQyQglfRCQjlPBFRDJCCV9EJCOU8EVEMkIJX0QkI5TwRUQyQglfRCQjlPBFRDIilnvaiojE6rJuzZ6vq009EkZn+CIiGaGELyKSEUr4IiIZoYQvIpIRSvgiIhmhhC8ikhG6LFOi06VvIqmiM3wRkYxQwhcRyQglfBGRjFDCFxHJCCV8EZGMUMIXEckIJXwRkYyoSsI3s1vN7F0zW1xguZnZtWb2hpm9bGYHVKNcEUmRy7ptmaQmqnWGfxtwTJHlxwJ7hNMk4JdVKldEREpUlV/auvuTZjawyConAne4uwPPmFl3M+vj7qsiFahfeIqIlC2uPvxdgXdynjeE87ZiZpPMbL6ZzV+zZk1MVRMRyYZEfWnr7je5+zB3H9a7d+9aV0dEpE2JK+GvBPrlPO8bzhMRkZjElfBnA+PCq3UOAdZF7r8XEZFIqvKlrZnNAI4AeplZA3ApUAfg7r8CHgSOA94APgLOrEa5IiJSumpdpfP1FpY7cF41yhIRkWgS9aWtiIi0HiV8EZGMUMIXEckI3dO2FPplr4i0ATrDFxHJiGyd4etMXUQyLFsJX0SkNaTkZFJdOiIiGaGELyKSEerSEZFky+0uSWhXSVroDF9EJCOU8EVEMkIJX0QkI5TwRUQyQglfRCQjdJWOSFbp6pfMUcIXqbaoiVQJWFqZunRERDJCZ/hJlJJxOUQkXXSGLyKSEUr4IiIZoYQvIpIRSvgiIhmhhC8ikhFK+CIiGaGELyKSEboOX6QQ/fJV2hgl/NakH1CJSIKoS0dEJCOU8EVEMkIJX0QkI5TwRUQyoioJ38yOMbPXzOwNM5uSZ/kEM1tjZgvD6exqlCsiIqWr+CodM2sPXA+MAhqA581strsvabbq3e5+fqXliZRNl1eKANU5wz8IeMPd/+zunwJ3ASdWYbsiIlJF1Uj4uwLv5DxvCOc1d7KZvWxm95lZvyqUKyIiZYjrh1cPADPc/RMzOwe4HTiq+UpmNgmYBNC/f/+YqiYiEmrjP5asxhn+SiD3jL1vOK+Ju69190/CpzcDB+bbkLvf5O7D3H1Y7969q1A1ERFpVI2E/zywh5kNMrOOwBhgdu4KZtYn52k9sLQK5YqISBkq7tJx901mdj7wCNAeuNXdXzGzy4H57j4buMDM6oFNwPvAhErLFRHJrIhdT1Xpw3f3B4EHm837Yc7ji4CLqlGWiIhEo9EyRdJOvzOQEmloBRGRjFDCFxHJCHXpSHqo60KkIjrDFxHJCCV8EZGMUJeOiJRHXWuppTN8EZGMUMIXEckIJXwRkYxQwhcRyQglfBGRjFDCFxHJCCV8EZGMUMIXEckIJXwRkYxQwhcRyQglfBGRjFDCFxHJCCV8EZGMUMIXEckIDY8sIlIruUNNQ6sPN60zfBGRjFDCFxHJCCV8EZGMUMIXEckIJXwRkYxQwhcRyQglfBGRjFDCFxHJCCV8EZGMUMIXEckIJXwRkYyoSsI3s2PM7DUze8PMpuRZ/hkzuztc/qyZDaxGuSIiUrqKE76ZtQeuB44FvgB83cy+0Gy1icDf3H134OfAVZWWKyIi5anGGf5BwBvu/md3/xS4Czix2TonAreHj+8DRpqZVaFsEREpUTWGR94VeCfneQNwcKF13H2Tma0DegLv5a5kZpOASQD9+/ffagMDp/yx6fGKqaUPIaq49MblxihOceXEJfWYrkVcrkR9aevuN7n7MHcf1rt371pXR0SkTalGwl8J9Mt53jecl3cdM+sAdAPWVqFsEREpUTUS/vPAHmY2yMw6AmOA2c3WmQ2MDx+fAjzm7l6FskVEpEQV9+GHffLnA48A7YFb3f0VM7scmO/us4FbgOlm9gbwPsGHgoiIxKgq97R19weBB5vN+2HO4w3AqdUoS0REoknUl7YiItJ6qnKGLyKSViumfrXWVYiNzvBFRDJCCV9EJCOU8EVEMkJ9+CIVylIfsKSbEr5IjeiDIptq+bqrS0dEJCN0hi+JpTNgkepSwpc2Rx8UyaDXIXnUpSMikhFK+CIiGaGELyKSEerDFwmpzznd9Pq1TGf4IiIZoTN8ESlKZ85th87wRUQyQglfRCQj1KUjIomiLqTWozN8EZGM0Bl+ATrLEJG2Rgm/yqJ+UOgDRkSKqUaOUMIXEYkgjSdpbT7hp/FFERFpDW0+4Yu0Neo2lKh0lY6ISEboDD+j4j7b09mlSO3pDF9EJCOU8EVEMkJdOimnrhIRKZXO8EVEMkJn+FIW/Uchkl4VneGbWQ8ze9TMloV/dyyw3mYzWxhOsyspU0REoqm0S2cKMNfd9wDmhs/z+djdh4ZTfYVliohIBJUm/BOB28PHtwNfq3B7IiLSSipN+Du7+6rw8V+BnQus18nM5pvZM2ZW8EPBzCaF681fs2ZNhVUTEZFcLX5pa2b/BeySZ9H/zX3i7m5mXmAzA9x9pZl9DnjMzBa5+/LmK7n7TcBNAMOGDSu0LRERiaDFhO/uXym0zMxWm1kfd19lZn2AdwtsY2X4989m9gSwP7BNwhcRkdZTaZfObGB8+Hg8MKv5Cma2o5l9JnzcC/gisKTCckVEpEyVJvypwCgzWwZ8JXyOmQ0zs5vDdQYD883sJeBxYKq7K+GLiMSsoh9euftaYGSe+fOBs8PHTwNDKilHREQqp6EVREQyIjVDK+gn/SIildEZvohIRijhi4hkhBK+iEhGKOGLiGSEEr6ISEYo4YuIZIQSvohIRijhi4hkhBK+iEhGmHsyh503szXAWwUW9wLei7BZxSlOcdmIS0MdWytugLv3zrvE3VM3AfMVpzjFKS4JZaUpTl06IiIZoYQvIpIRaU34NylOcYpTXELKSk1cYr+0FRGR6krrGb6IiJRJCV9EJCOU8EVEMiI1tzhsZGbbAxvcfXMZMTsCnwU+Bla4+z9as7xy48xsBHAGcBjQJ6znYuCPwG/cfV0L8ZH2L852KbesqG1SQVxfYEwY99lmcQ+V0jZJPlbULsXfQ3HVMycm9pwEJP+HVwT/hfwTwQv3LvBO+HcJMA3YvUBcN+BiYBHwGvAUMD+Mvxc4ssrlRY17CLgFqCc4ADoAOwAHAN8DngDqq7h/sbVLBWVFbZOocb8G5gAXAIcCuwP7AKOB64CngcNTfKyoXfLvX9z1jDUn5d1WuQk47gmYB1wC7Au0y5nfAzgZ+B1wRp64R4FvAN3zLDsQuAaYWMXyosb1KqENtlmngv2LrV0qKCtqm0SN26eFmI753lQpOlbULvn3L+56xpqT8k2JvyzTzOrcfWOl67R2eXHXM25tff+i0LGSX1rapa3XM5/Ef2lbbCfMbIeW1ikQt1e1y2ulei4qZ/2cuIL7FzWu2vtXQR2jtknUuIcKLWsjx0pm2yVh9YzlvZe6L22bWQL0jxA3J2Jc1PIKxpnZ6AIxBuwSoSyIvn9xtkvBsqK2SQVxBxSJG1oorgWJOVbULpFUvZ4tiOW9l/iEb2bfLbSI4IuZQnHXFonr3grlRYoD7gbuBPL1rXUqUl7U/YutXaKWRcQ2qSDueYJ+UsuzLPXHSgVxbbpd4q5n3Dkpb0AK+vA3EHwTvSnP4n9297wNZWYfEnxD/0mexVe7e68qlxc1bgEw3t0X51n2jrv3KxAXdf9ia5cKyoraJlHjFgMnufuyMuPScqyoXZJRz1hzUl6lfLNby4ng0q8DCyx7p0jcY8ChBZa92QrlRY07DOhfYNmwVti/2NqlgrKitknUuFOAzxdY9rU2cKyoXZJRz1hzUr4pDWf4nwfWuvs2d3cxs53dfXWBuB4EP074KKbyIsVFVcH+xdYuUctKi7QcK3FLS7vEXc+4c1LebSU94YuISHUk/rJMERGpDiV8EZGMUMJPKDM70cwOrnU9kiRqm1QQN8zMPltuXNzULvnpPbStxF+HX4iZTQbWAr9z93yXKxWKuxJYB9zs7mtjKC9SHHAwMMTMOrj7sWWUF3X/YmuXqGURsU0qiPs2sK+Zve7up5calJZjpYK4Nt0ucdcz1vdeWr+0NbPzgL2AAe5eX0bc14DdgP3cfVwM5UWKi6qC/YutXaKWVStm1sXdPyxj/VQcK5Vqq+0Sdz1jfe+lNeG3JRaMo3EisGs4ayUw292X1q5WtRW1TSqI6wYc0yzuEXf/INoetA61S356D5UmFQnfzI4GvsbWL+Ysd3+4SEwHYCJwEsEY2U1xwC1efECissuroJ7/CnwduAtoCGc33nTiLnefWs39i7NdKqhj1DaJGjcOuJRgPJOVOXGjgB+5+x354sLYNBwrapdk1DP2nLTNdpKe8M3sGmBP4A62fjHHAcvc/cICcTOAD4Dbm8WNB3oU6nusoLyoca8Dezd/sc2sI/CKu+9R5f2LrV0qKCtqm0SNew04uPlZqwV3JXrW3fcsEJeWY0Xtkox6xpqT8vIyfpZbiwl4vcB8C3e2rLioy6KWV0LcqwR9cM3nDwBei2v/WqNdKigraptEbkugW5753VrpNY/7WFG7JKOeseakfFMartLZYGbD3f35ZvOHAxuKxL1vZqcSfIP9DwAzawecCvytFcqLGvcdYK6ZLSO4dRkEw53uDpxfJC7q/sXZLlHLitomUeN+ArxgZnOaxY0CflwkLi3HitolGfWMOydtIw1dOgcAvwS6sOXfmX4ElzGd5+4LCsQNBK4CjmJLY3YHHgemuPubVS4vUlwY2w44iK375573IjcprmD/osaVvX95yjKCs8OiZYWxZbdJhXE7Akez7ZeTBd+IaTlWKoxrs+0Sdz3jzkl5t5X0hN/IzHYh58V097+WEdsTwMu7xjVSeVHiwoMVd/9H2O+4D8Gd7N8vscyy9y9qXAXtEqmOebazg7uvr2QbJZbTo4z2j+1YKbKtTLdLJe+hWrx+cb73tlJO/0/SJmCviHGjWqs8oC7PvII3WSb45n01sIrgsrJngbkEn+QntMb+AV2B3fLM37eFuF2AXcLHvYHRwBfKqNegMCbS6xZu4+0iy/YFniH4t/4mYMecZc8VifsisBR4heDHOo8Cy8PtjEjKsaJ2Kbh+a7yH4n79Wi0nbbV+lEKSMhU7yOOOA44MD7D3CC5jG5iz7IUicS+GiXQQ8L+E448TfOE0vxXqeRrwF2Bh+EYeXmI9zwHeBFYA54ZvqluA14CJBWJm5jw+MYz/NcGXgROKlPXdAtP3gPeLxD1FcM14d+Bfwv3brbGdi8Q9BwwBRoSv35fC+QcA/52gY0Xtkoz3UKR6Ri2vmnGJ/9LWot8WbHaRuJ7VLg/4KXC0u79iZqcAj5rZN9z9mTC2IA//NTOzt939tXDeW43/phaoZ6T9Ay4muJnCKjM7CJhuZhe5+/0t1PN8YG9gO+AtYHd3/2vYx/s4QfJvbkDO438FjnL3N82sF8EZ2G0FyrqSwnf4KTb+Uxffcl3yzyy4E9LDZvYN8t/+rlGduy8CMLM17v4UgLu/YGbbFQqqwbGidikg4nso1nrWICdtI/EJHziTwrcF+3qRuMOAM4Dm/ZpG8OVOtcvr6O6vALj7fWa2FPi9BT8KKfpFiZm18+Bb+7Ny5rUHOhYJi7p/7d19VVjP58zsSOAPZtavhXpu9ODGDR+Z2fLGN5i7/83MCsXlzu/g4ZdS7v6emf2jSFkvEPx3kO+L4LOLxGFm3dx9XVjO42Z2MvA7oEeRsNykcFGzZcVeg7iPFbVLARHfQ3HXM+6ctK0o/0bEORH9tmAPAUcWWPZkK5Q3n7B/O2deX4Kukw+LxA0HOuWZPxA4oxX272ma9d8TfPs/F/ikSNwCwj5LoG/O/E7ASwViNhP8i/0h8CnQJ5zfEXi5SFmfB3oXWLZzkbh/Ag7JM78/8J9F4uqBznnm7wb8IEHHitolf1zU91Dc9Yw1J+WbEn+VjsV8m7yo5ZnZV4A17v5Ss/ndCS6d+kmJZeMlXgURhZntB/zd3d9oNr8OOM3d7ywQ1x/4izcblc/MdgUGu/t/lVGH7mHM/5S9AwlSy2MlydLSLm29nnm3lfSEnyuOhFiN8sqJCxPpT4GRBD+7NoKraB4juDZ3RXm1Tj8LBuy6iODqi50I/k1+l2DMkaleYOCuuOPipnbJz8yO8fA7ijDpXk1w1r8Y+Gcv4Z6vcbzXq6HS8hJ/AxQz629md5nZGoIrQ54zs3fDeQOLxOX25fU1s7lm9oGZPW1meccAqbC8SHHA3cD9BP8i7uHuuwN9gJkEg0FVe/9ii4taFnAPwQ9TjnD3Hu7ek+DKiL+FyxIRZ2bH5DzuZma3mNnLZvZbM9u52nFx71/UuBq0y5U5j39GcHnmCcDzwI1Fyov1vR53TsqrnP6fWkzA/wCnE3zZ2DivPcFIeM8UiXsh5/E9wCSCD7iTgLmtUF7UuGJjbxRbFnX/YouroKxi458kaVnu/t0MXEFwZdI/k3NJahXjkrTvSWqX3LiFzZYtLBIX93s91pyUd1vlrFyLqYWkV2pCbH4QFLv2OGp5UZfdBdxA8MOWz4bTweG8e1ph/2KLq6CsOcAPyPkiEtiZ4NLO/0pQXNREEzVO7ZI/roEtv0f4M2FXdbis2MUBcb/XY81J+aY0XJa5wMxuIBhStHFgpH4EQ4q+WCSurwXXrxrQ28zqfMvwqXWtUF7UuHEEY2T/iC0/m24AHiD/te2Nou5fnHFRyzodmALMM7OdwnmrgdkEPxxLStxOZvZdwu9dzMw8fBdSvLs0apzaJb//JLjSDIL3Xy9gjQVDESwsEhf3ez3unLSNxH9pa8G4GBPZ+m42TQnR3fNdm4qZjW82a7YH14zvAlzg7hdXubxIcVFVsH+xxUUtKy3M7NJms25w98ZE81MvcLu6qHFpkZZ2KfCeXUnwgVbue72UuGrmpBbLy7utpCf8LLDgbjZ9Cf5Nfitn/lnufmvtalYbZmYEQ8Y6cB/B6IInEox7/isPh5atdVwSmNlj7n5U0uPilpZ6xi3VCd/Mfujul1c7LmoCjhJnwR3rv0TwK8oTgGvc/bpw2QvufkAJ5c31nMs3y6hnq8dFjLmB4DLAjgQ/3PoMwdnMV4HVXviOQnHHxf3B9HLzWQR3QmocSmDfJMQV2FarfcBUsH+93P29nOdnEPzidTHBD9LyJscaxJ0EzHP3982sN8GVSAcAS9zZbUAAAA07SURBVIDvuXtDvri8yunwT9pE6wyCdiXwJHANwciA385ZVmwAp6hxiwiGHYBgXIwHgZ97y1/k/L+I5cUWV0FZi8K/dcBagp+yQzAUSLEv4eKOu4EgYc8GfgPcC3yD4Iv4/2iFuMb19yK4emUgQZ/uAPLc8amGcS83mxYRDAvwcgvtGTUuaj1zv0T9N+ARgn7xewnfgwmJW5Lz+G6Cq5b6AhOARwvF5d1WOSvXYiI448o3fQhsaoW4qAk4atzSZs/bE3xZey/B/TiTUs+y4yoo68Wcxw83W1bsao2442L9gAnXOYngQ7Q+fP7nEt9HscUR8wdMBfXMfd1fALbPeV0WJSjutZzHC0o9PvNNif/hFcGvT/dw967Npi4EP7CodlwHD4cP8OCXhCcQXDFwL8UHYooat9zMvtz4xN03u/tEgn9HByeonlHiopb1VzPbIYzL/THOLgRj8iQlrnHfNhLcXenT8PkmoFi/f9Q4PBjV9FjgCDObRfF2rEmcu9cTDMx2E7CfB115G939Lc/p6qxWXAX7t52Z7W9mBxJc4/73cFsbCcaBSkrcE2Z2uQUjlD4RdvFgwcCH60rYzy3K+XSoxUTw44uDCiy7qhXi/gB8ucD2/tEKcdsB2xVYtmvO471rXM+y46KWVaQO2wM7FWqTuOMIBsPaIU/cLhS/sUikuDzr7wd8K8/8ovsXV1zYfv9OMBRDQ5mvc9lx5daTYFjv3KlxcL+eFBlHvwZxdcBlwNvh9A+CnorfAv3LaptyGzOpU6lv4hIOgqgJOFJcGfV8oRrlxRkXd5skJY6YP5iS3i7lflBUGleF1689OaOElvH6tXocwT2hexZY1mJ5ZTdaUqekHeStEFewzzth9Sw7rgZtEndcWo4VtUsy6tlqcWnowy9V0TvitIE4j7m8OOPibpO449JyrKhdqlte4uLaUsJPy0EeNS6qNOxf3G0St7QcK3FLS7uk4T1UUlxbSvhtXbGrRbIqapvEHRc3tUt+aalnq0nD4GmlSstBXjDOzEYT/OrWgac8uNQMAHc/pNrlJSiu6m0Sd1wLEnOsqF2qW17a4lI1tEKxFzPtceHP+3cHZoSzTgeWu/t5Sapn1Lg42yTuuDA28ceK2iUZ9axFXFN8WhJ+Wg7yCuJeJbjPq4fP2xH80rbYj69SsX81aJO449JyrKhdklHP2D94m0S5/KcWE8HAUrk3NmhHs2EJUh73B3J+Pk7wc/IHEljPsuNq0CZxx6XlWFG7JKOescblTmn60vYNoH/O837hvLYS1wVYamZPmNnjBCPhdTWz2WY2O0H1jBIXd5vEHZeWY0Xtkox6xh3XJE1dOvMI7kT/HEH/1UHAfMKxJDwYhyPNcV8uvPfg7vMSUs+y42rQJnHHpeVYUbsko56xxm21jRQl/LQc5JHiokrD/sXdJnFLy7ESt7S0SxreQ5XEbbWNtCT8ts7MPmTLDyc6EgyY9Hd371q7WtVW1DaJOy5uapf80lLPWkrNdfhpOcijxnkwbHPjNozgDkgtXjechv2Lu03ijkvLsaJ2SUY9a/rBW843vEmZAAO+Bkxti3E58WUN9pSG/Yu7TeKOS8uxonZJRj3jjkt1l46Zveju+7eFOAt+UNGoHTCMYCz5Ea1RXq3jWrNN4o4rsK3EHStql+qWl8a4NHXp5HsxN7SVOIK7QTXaBKwg+Jc0UfWMEhd3m8Qdl5ZjJe64tLRLGt5DlcTlSk3CJyUHedQ4dz+zhG1XrbyY42Jtk7jjSMmxonapbnkpitsiSr+YpupPBHehvx94N5x+B/Stdb3S2CZxx6ldkjGlpZ41baNaV6C1X8wUxT0KnEnwX1cHYALwaALrWXZcDdok7ri0HCtql2TUs2YfvCWvWOspRQd51LiFpcxLQD3LjqtBm8Qdl5ZjRe2SjHrGGrfVNspZuZZTig7yqHFzgTMIbmjcPnw8N4H1LDuuBm0Sd1xajhW1SzLqGWtc7pSmwdPWmtkZZtY+nM4A1rahuLOA04C/AquAUwg+zZNWzyhxcbdJ3HFpOVbULsmoZ9xxW5Tz6VDLiWCo09nAGoL+q5lA/7YQR3A2cmdbbZc42yTuuLQcK2qXZNSzFnFbbSNKA8U9peUgr/CgewromOR6RomLu03ijkvLsaJ2SUY9a9EuuVMqrsN3981mNsDMOrp7yfd7TEtc6M/Af1swbvffc7b570mpZ5S4uNsk7ri0HCtxx6WlXdLwHqokrrlUJPxQ4g/yCuOWh1M7ghs5lCoN+xd3m8Qdl5ZjRe2SjHrGHdckTQk/LQd5pDh3/1EZZVRcXsxxsbZJ3HGk5FhRu1S3vBTFNUn14GltiZk9wJahTxutI7ijzY3uXtaYGW1B1DaJOy5uapf80lLPWkpNwk/LQV5B3H8Avdn6jvT/G26rq7t/IyH1LDuuBm0Sd1xajhW1SzLqWbMP3jQl/LQc5FHjnnf34fnmmdkr7r53QupZdlwN2iTuuLQcK2qXZNQz1riteIWX+cQ1Ac8Xmge80gbilpJzTS3B3emXho8L3sQhDftXgzaJOy4tx4raJRn1jDUud0rTl7Y7mFl/d38bwMz6AzuEy4pdppSWuO8BT5nZcoK72QwCJpvZ9sDtCapnlLi42yTuuLQcK2qXZNQz7rgtSvlUSMIEHAe8DTwOPAG8BXwV2B74TtrjwtjPAPuFU6e20i5xt0nccWk5VtQuyahnLdqlcUpNHz6AmX0G2Ct8+pqX+K17GuJs67vZNFoHLHL3d5NSz6hxcbZJ3HFhbOKPFbVLMupZi7im+LQk/LQc5BXE/REYATxG8O/oEcACgn9LL3f36QmpZ9lxNWiTuOPScqyoXZJRz9g/eJuU+i9drSfgj8D7wH0EA/+vBeYAy4BvtIG4R4Cdc57vHM7rASxOUD3LjqtBm8Qdl5ZjRe2SjHrGGrfVNkpZKQlTig7yqHFLmj23xnkUv8Ig8ftXgzaJOy4tx4raJRn1jDUud0rTVTr93H11zvN3w3nvm9nGNhD3hJn9Abg3fH5yOG974IME1TNKXNxtEndcWo4VtUsy6hl3XJM0Jfy0HORR484DRgNfCp/fAfzOg4/yIxNUzyhxcbdJ3HFpOVbULsmoZ9xxTdL0pa2x9Yv532x5MVMbZ2ZWyjaLxcdRz0ri4mqTuONyl5HgY0Xtkox61ipuK6X0+9RyIvxQKnedFMU9AXybZneuAToCRxH8YGRCAupZdlwN2iTuuLQcK2qXZNQz1ri865WyUi2nFB3kUeM6AZMJPq3/AiwB3iT4UcV/AvuntV1q0CZxx6XlWFG7JKOescblmxLfpWNmnQhuTjyW4HraD4DtCMaEngPc4O4vpjWu2TbqgF7Ax+5etE8uDfsXd5vEHZeWYyXuuLS0SxreQ9Vql6ZtJT3h50ryQV6NuKjSsH9xt0nc0nKsxC0t7ZKG91AlcU3xaUr4IiISXbtaV0BEROKhhC8ikhFK+JIZZvagmXVvYZ31BebfZmantE7NwMwuM7N/aa3ti0C6fmkrEkn4gxVz9+NqXReRWtIZvqSGmU01s/Nynl9mZv9mZnPN7AUzW2RmJ4bLBprZa2Z2B7AY6GdmK8ysV7h8ppktMLNXzGxSs3J+Hs6fa2a989TjQDObF8Y/YmZ9mi3vZmZvmVm78Pn2ZvaOmdWZ2TfN7Hkze8nMfmdmnfNs/wkzGxY+7mVmK8LH7c1sWhj/spmdU2GTSsYo4Uua3A2clvP8NIIfnZzk7gcQjJdydXhGD7AHwTXKe7v7W822dZa7HwgMAy4ws57h/O2B+R7c8HoecGluUHhZ3HXAKWH8rcBPctdx93XAQuDL4azjgUfcfSPwe3cf7u77EdyDdWIZ+z8RWOfBjbqHA980s0FlxEvGqUtHUsPdXzSznczss0Bv4G/AX4Gfm9nhwD+AXQmGjQV4y92fKbC5C8zspPBxP4IPh7XhNu4O5/8G+H2zuM8D+wCPhp8r7YFVebZ/N3A6we3oxgA3hPP3MbMrgO4E9yN9pIRdb/R/gH1zvkvoFtb7zTK2IRmmhC9pcy9wCrALQVIdS5D8D3T3jWH3R6dw3b/n24CZHQF8BRjh7h+Z2RM5Mc01/6GKAa+4+4gW6jkbuNLMegAHEtyFCeA24Gvu/pKZTSC4K1Nzm9jy33duvQz4truX8yEh0kRdOpI2dxOcMZ9CkPy7Ae+Gyf5IYEAJ2+gG/C1M9nsBh+QsaxduG+CfgKeaxb4G9DazERB08ZjZ3uHj883sfAB3Xw88D/wH8Ad33xzGdwFWhV1DYwvUbwXBhwQ5dYHgv4Fzw1jMbE8LhsYVKYkSvqSKu79CkDRXuvsq4E5gmJktAsYBr5awmYeBDma2FJgK5Hb7/B04yMwWEwxMdXmz8j8lSMJXmdlLBH31h4aL9yLoFmp0N3AGW7qIAC4BniUY6KtQXX9GkNhfJPgZfaObCQYGeyGs343ov3Qpg4ZWEKkSC25OMTr8UBBJHCV8EZGMUJeOiEhGKOGLiGSEEr6ISEYo4YuIZIQSvohIRijhi4hkxP8HKt7slSr0YY8AAAAASUVORK5CYII=\n","text/plain":["<Figure size 432x288 with 1 Axes>"]},"metadata":{"tags":[],"needs_background":"light"}}]},{"cell_type":"markdown","metadata":{"id":"QZEflJLQDtVT"},"source":["**RetEarly2020**\n","\n","- Among all the industries, the telecommunication services sector (`ggroup_5010`) has the highest return rate in 2020, indicating the magnitude telecommunication sector impacted by the pandemic is the least. However, the negative change of return suggests the telecommunication services sector faced a loss in early 2020. We can infer that during the pandemic, people still rely on communication and require telecommunication services.\n","- Compared to the other industries, the energy sector (`ggroup_1010`) has the lowest returns in 2020. During the pandemic, people were confined to their homes and restricted to travel, resulting in a declined oil demand and a decrease in returns, which made the energy sector the sector impacted by the pandemic the worst.\n","\n","**RetLate2020**\n","- For late 2020, the consumer discretionary sector (`ggroup_2510`) has the highest return compared to the other industries. The pandemic caused a consumption shock due to the lockdown and forced shutdown across several industries. However, during recovering from the pandemic in 2020, people may have been accumulating more savings while forced to cut back on spending from lockdowns and travel restrictions. Thus, in late 2020, the consumer discretionary sector may have the highest bounce in return.\n","- In contrast, the utilities sector (`ggroup_5510`) has the lowest return growth among all the industries. If we compare to early 2020, the magnitude utilities industry affected by the pandemic was less than the other industries. We may infer that the pandemic comparatively less affected the sector. After the COVID-19 virus has been controlled, the extent it recovered from the pandemic would also be less than the other sectors. On the other hand, the utilities were affected by the stopped commercial activities and restrictions on public transport; power demand might take time to recover from the pandemic."]},{"cell_type":"markdown","metadata":{"id":"GDQilylQDF1W"},"source":["#### ii. Fixed effect regressions\n","Run fixed effect regressions (using industry fixed effects) to explain both *ReturnEarly2020* and *ReturnLate2020:*\n","\n","1. You may wish to review the lecture notes from Lecture #3 and use the Colab Notebook: Fixed_Effect_Regression_for_BA870_Lecture_3.ipynb that is posted on QTools in Lecture #3 folder.\n","2. What is the explanatory power of these regressions using different types of industry variables. Is this high or low?Explain.\n","\n","Combine the industry fixed effects with other explanatory variables (see Assignment #3 and/or the financial ratios described in OPTION (3) below).\n","   - Run combined regressions with both industry and other “X” variables."]},{"cell_type":"markdown","metadata":{"id":"YQdyIvYdVzit"},"source":["**RetEarly2020**"]},{"cell_type":"code","metadata":{"id":"iNR9ZwhIDGzO","colab":{"base_uri":"https://localhost:8080/"},"executionInfo":{"status":"ok","timestamp":1625713113140,"user_tz":240,"elapsed":33,"user":{"displayName":"Tzu-Hua Huang","photoUrl":"https://lh3.googleusercontent.com/a-/AOh14GgbXOdZWwXGVTDuMHbeSme5C5oKAeF6I6ojY6qc=s64","userId":"01458545184256376706"}},"outputId":"6a3cce0b-cce7-490c-bc02-c53fa6d51d35"},"source":["# define x as a subset of original dataframe\n","x1_gics = df_gics.drop(columns=['TICKER', 'RetEarly2020', 'RetLate2020'])\n","# define y as a series (RetEarly2020)\n","y1_gics = df_gics['RetEarly2020']\n","\n","# run fixed effect regression\n","# pass x as a dataframe, while pass y as a series\n","model1_gics = sm.OLS(y1_gics, x1_gics).fit()\n","\n","# print out regression summary statistics\n","print(model1_gics.summary())"],"execution_count":241,"outputs":[{"output_type":"stream","text":["                            OLS Regression Results                            \n","==============================================================================\n","Dep. Variable:           RetEarly2020   R-squared:                       0.131\n","Model:                            OLS   Adj. R-squared:                  0.123\n","Method:                 Least Squares   F-statistic:                     17.25\n","Date:                Thu, 08 Jul 2021   Prob (F-statistic):           1.57e-64\n","Time:                        02:58:33   Log-Likelihood:                -501.12\n","No. Observations:                2665   AIC:                             1050.\n","Df Residuals:                    2641   BIC:                             1192.\n","Df Model:                          23                                         \n","Covariance Type:            nonrobust                                         \n","===============================================================================\n","                  coef    std err          t      P>|t|      [0.025      0.975]\n","-------------------------------------------------------------------------------\n","ggroup_1010    -0.5807      0.029    -20.189      0.000      -0.637      -0.524\n","ggroup_1510    -0.3509      0.027    -13.158      0.000      -0.403      -0.299\n","ggroup_2010    -0.3169      0.019    -16.455      0.000      -0.355      -0.279\n","ggroup_2020    -0.3325      0.032    -10.327      0.000      -0.396      -0.269\n","ggroup_2030    -0.3266      0.041     -8.027      0.000      -0.406      -0.247\n","ggroup_2510    -0.4295      0.054     -7.884      0.000      -0.536      -0.323\n","ggroup_2520    -0.3915      0.032    -12.085      0.000      -0.455      -0.328\n","ggroup_2530    -0.4482      0.031    -14.414      0.000      -0.509      -0.387\n","ggroup_2550    -0.3854      0.029    -13.267      0.000      -0.442      -0.328\n","ggroup_3010    -0.1759      0.063     -2.812      0.005      -0.298      -0.053\n","ggroup_3020    -0.1810      0.038     -4.779      0.000      -0.255      -0.107\n","ggroup_3030    -0.2310      0.060     -3.858      0.000      -0.348      -0.114\n","ggroup_3510    -0.1343      0.022     -6.021      0.000      -0.178      -0.091\n","ggroup_3520    -0.1410      0.017     -8.423      0.000      -0.174      -0.108\n","ggroup_4010    -0.3685      0.017    -21.610      0.000      -0.402      -0.335\n","ggroup_4020    -0.3596      0.026    -13.648      0.000      -0.411      -0.308\n","ggroup_4030    -0.2495      0.032     -7.794      0.000      -0.312      -0.187\n","ggroup_4510    -0.1765      0.022     -8.208      0.000      -0.219      -0.134\n","ggroup_4520    -0.2896      0.030     -9.519      0.000      -0.349      -0.230\n","ggroup_4530    -0.2265      0.037     -6.080      0.000      -0.300      -0.153\n","ggroup_5010    -0.0892      0.066     -1.360      0.174      -0.218       0.039\n","ggroup_5020    -0.3414      0.034    -10.079      0.000      -0.408      -0.275\n","ggroup_5510    -0.1479      0.035     -4.188      0.000      -0.217      -0.079\n","ggroup_6010    -0.3281      0.022    -14.839      0.000      -0.371      -0.285\n","==============================================================================\n","Omnibus:                     4272.607   Durbin-Watson:                   1.865\n","Prob(Omnibus):                  0.000   Jarque-Bera (JB):          4931809.804\n","Skew:                           9.949   Prob(JB):                         0.00\n","Kurtosis:                     212.805   Cond. No.                         3.92\n","==============================================================================\n","\n","Warnings:\n","[1] Standard Errors assume that the covariance matrix of the errors is correctly specified.\n"],"name":"stdout"}]},{"cell_type":"markdown","metadata":{"id":"latFf_noXdS8"},"source":["**R-squared & Adj. R-squared**\n","\n","- `R-squared`: 0.131\n","\n","Around 13% of the variation in `RetEarly2020` is explained by the industries that the companies belong to. We can infer that returns in early 2020 are described by its business model of the type of business that operates. So 13% of the variation which varies across all these companies is explained by the sector itself.\n","\n","- `Adjusted R-squared`: 0.123\n","\n","However, the Adjusted R-squared is less than the R-squared, which is 0.123. Since Adjusted R-squared considers different independent variables against the model, a decrease inA djusted R-squared indicates the additional variables doen't improve the model more than would be expected by chance.\n","\n","**t-stats**\n","\n","Among all the variables, except `ggroup_5010`, all the sector variables have large t-values, indicating the most significant difference compared to the other models. Their p-values are also far less than 0.05. We can reject the null hypothesis and conclude that the data provide enough evidence to reject the null hypothesis, suggesting that each variable is statistically significant. Data favor the hypothesis that there is a non-zero correlation between almost all the variables and `RetEarly2020`."]},{"cell_type":"markdown","metadata":{"id":"hwMO4Q-bXd4g"},"source":["**RetLate2020**"]},{"cell_type":"code","metadata":{"colab":{"base_uri":"https://localhost:8080/"},"id":"6hOWmtRMXd4g","executionInfo":{"status":"ok","timestamp":1625713113141,"user_tz":240,"elapsed":25,"user":{"displayName":"Tzu-Hua Huang","photoUrl":"https://lh3.googleusercontent.com/a-/AOh14GgbXOdZWwXGVTDuMHbeSme5C5oKAeF6I6ojY6qc=s64","userId":"01458545184256376706"}},"outputId":"3303e301-d201-418f-d729-9b40a0ca602b"},"source":["# define x as a subset of original dataframe\n","x2_gics = df_gics.drop(columns=['TICKER', 'RetEarly2020', 'RetLate2020'])\n","# define y as a series (RetLate2020)\n","y2_gics = df_gics['RetLate2020']\n","\n","# run fixed effect regression\n","# pass x as a dataframe, while pass y as a series\n","model2_gics = sm.OLS(y2_gics, x2_gics).fit()\n","\n","# print out regression summary statistics\n","print(model2_gics.summary())"],"execution_count":242,"outputs":[{"output_type":"stream","text":["                            OLS Regression Results                            \n","==============================================================================\n","Dep. Variable:            RetLate2020   R-squared:                       0.110\n","Model:                            OLS   Adj. R-squared:                  0.102\n","Method:                 Least Squares   F-statistic:                     14.17\n","Date:                Thu, 08 Jul 2021   Prob (F-statistic):           8.28e-52\n","Time:                        02:58:33   Log-Likelihood:                -3816.5\n","No. Observations:                2665   AIC:                             7681.\n","Df Residuals:                    2641   BIC:                             7822.\n","Df Model:                          23                                         \n","Covariance Type:            nonrobust                                         \n","===============================================================================\n","                  coef    std err          t      P>|t|      [0.025      0.975]\n","-------------------------------------------------------------------------------\n","ggroup_1010     1.0175      0.100     10.195      0.000       0.822       1.213\n","ggroup_1510     0.8732      0.093      9.437      0.000       0.692       1.055\n","ggroup_2010     0.8981      0.067     13.440      0.000       0.767       1.029\n","ggroup_2020     0.6055      0.112      5.419      0.000       0.386       0.825\n","ggroup_2030     0.7097      0.141      5.028      0.000       0.433       0.987\n","ggroup_2510     1.8114      0.189      9.583      0.000       1.441       2.182\n","ggroup_2520     1.5944      0.112     14.184      0.000       1.374       1.815\n","ggroup_2530     1.1311      0.108     10.484      0.000       0.920       1.343\n","ggroup_2550     1.7803      0.101     17.665      0.000       1.583       1.978\n","ggroup_3010     0.4041      0.217      1.862      0.063      -0.021       0.830\n","ggroup_3020     0.5335      0.131      4.060      0.000       0.276       0.791\n","ggroup_3030     0.6115      0.208      2.943      0.003       0.204       1.019\n","ggroup_3510     0.7397      0.077      9.559      0.000       0.588       0.891\n","ggroup_3520     0.7397      0.058     12.734      0.000       0.626       0.854\n","ggroup_4010     0.4338      0.059      7.333      0.000       0.318       0.550\n","ggroup_4020     0.7456      0.091      8.157      0.000       0.566       0.925\n","ggroup_4030     0.3584      0.111      3.228      0.001       0.141       0.576\n","ggroup_4510     1.0355      0.075     13.875      0.000       0.889       1.182\n","ggroup_4520     0.8171      0.106      7.742      0.000       0.610       1.024\n","ggroup_4530     1.2186      0.129      9.427      0.000       0.965       1.472\n","ggroup_5010     0.4901      0.228      2.153      0.031       0.044       0.936\n","ggroup_5020     0.8547      0.118      7.272      0.000       0.624       1.085\n","ggroup_5510     0.1812      0.123      1.479      0.139      -0.059       0.421\n","ggroup_6010     0.4567      0.077      5.952      0.000       0.306       0.607\n","==============================================================================\n","Omnibus:                     2894.744   Durbin-Watson:                   1.965\n","Prob(Omnibus):                  0.000   Jarque-Bera (JB):           349263.463\n","Skew:                           5.263   Prob(JB):                         0.00\n","Kurtosis:                      58.087   Cond. No.                         3.92\n","==============================================================================\n","\n","Warnings:\n","[1] Standard Errors assume that the covariance matrix of the errors is correctly specified.\n"],"name":"stdout"}]},{"cell_type":"markdown","metadata":{"id":"zIF8vVUiXr3R"},"source":["**R-squared & Adj. R-squared**\n","\n","- `R-squared`: 0.110\n","\n","  Around 11% of the variation in `RetLate2020` is explained by the industries that the companies belong to. We can infer that returns in late 2020 are described by its business model of the type of business that operates. So 11% of the variation which varies across all these companies is explained by the sector itself.\n","\n","- `Adjusted R-squared`: 0.102\n","\n","  However, the Adjusted R-squared decreases compared to the R-squared, which is 0.102, indicating that the increase in the variables in the model doesn't enhance the explanatory power of the model. \n","\n","**t-stats**\n","\n","Compared to the regression model for `RetEarly2020`, there are more variables have large p-values. In general, most of the variables have large t-values, and their p-values are also far less than 0.05. We can reject the null hypothesis and conclude that the data provide enough evidence to reject the null hypothesis, suggesting that each variable is statistically significant. Data favor the hypothesis that there is a non-zero correlation between almost all the variables and `RetLate2020`. \n","\n","On the other hand, sector variables `ggroup_3010` and `ggroup_5510` have p-values greater than 0.05. We fail to reject the null, suggesting that the variables are not statistically significant."]},{"cell_type":"markdown","metadata":{"id":"oHAmJJvkqTls"},"source":["### 3.b NAICS"]},{"cell_type":"markdown","metadata":{"id":"57ZpJDrcqTlt"},"source":["#### i. Average stock returns\n","What are the average stock returns for each industry classification (23 industry groups)?\n","\n","1. Do the highest and lowest return industries make economic sense for early and late 2020? Explain in detail."]},{"cell_type":"code","metadata":{"colab":{"base_uri":"https://localhost:8080/","height":793},"id":"LdLTHDbaqTlv","executionInfo":{"status":"ok","timestamp":1625713113282,"user_tz":240,"elapsed":148,"user":{"displayName":"Tzu-Hua Huang","photoUrl":"https://lh3.googleusercontent.com/a-/AOh14GgbXOdZWwXGVTDuMHbeSme5C5oKAeF6I6ojY6qc=s64","userId":"01458545184256376706"}},"outputId":"3adef798-83b1-4a99-fdde-84ab7d833c6d"},"source":["df_naics_AvgRet = df_naics.drop(columns=['TICKER'])\n","df_naics_AvgRet = df_naics_AvgRet.melt(id_vars=['RetEarly2020', 'RetLate2020']).groupby(['variable', 'value']).mean()\n","df_naics_AvgRet.query('value == 1')"],"execution_count":243,"outputs":[{"output_type":"execute_result","data":{"text/html":["<div>\n","<style scoped>\n","    .dataframe tbody tr th:only-of-type {\n","        vertical-align: middle;\n","    }\n","\n","    .dataframe tbody tr th {\n","        vertical-align: top;\n","    }\n","\n","    .dataframe thead th {\n","        text-align: right;\n","    }\n","</style>\n","<table border=\"1\" class=\"dataframe\">\n","  <thead>\n","    <tr style=\"text-align: right;\">\n","      <th></th>\n","      <th></th>\n","      <th>RetEarly2020</th>\n","      <th>RetLate2020</th>\n","    </tr>\n","    <tr>\n","      <th>variable</th>\n","      <th>value</th>\n","      <th></th>\n","      <th></th>\n","    </tr>\n","  </thead>\n","  <tbody>\n","    <tr>\n","      <th>Sector_naics_11</th>\n","      <th>1</th>\n","      <td>-0.158294</td>\n","      <td>0.123019</td>\n","    </tr>\n","    <tr>\n","      <th>Sector_naics_21</th>\n","      <th>1</th>\n","      <td>-0.597364</td>\n","      <td>1.231517</td>\n","    </tr>\n","    <tr>\n","      <th>Sector_naics_22</th>\n","      <th>1</th>\n","      <td>-0.155907</td>\n","      <td>0.191994</td>\n","    </tr>\n","    <tr>\n","      <th>Sector_naics_23</th>\n","      <th>1</th>\n","      <td>-0.371588</td>\n","      <td>1.090058</td>\n","    </tr>\n","    <tr>\n","      <th>Sector_naics_31</th>\n","      <th>1</th>\n","      <td>-0.281907</td>\n","      <td>0.775376</td>\n","    </tr>\n","    <tr>\n","      <th>Sector_naics_32</th>\n","      <th>1</th>\n","      <td>-0.182609</td>\n","      <td>0.760634</td>\n","    </tr>\n","    <tr>\n","      <th>Sector_naics_33</th>\n","      <th>1</th>\n","      <td>-0.293057</td>\n","      <td>1.022835</td>\n","    </tr>\n","    <tr>\n","      <th>Sector_naics_42</th>\n","      <th>1</th>\n","      <td>-0.320072</td>\n","      <td>0.755734</td>\n","    </tr>\n","    <tr>\n","      <th>Sector_naics_44</th>\n","      <th>1</th>\n","      <td>-0.388449</td>\n","      <td>1.527506</td>\n","    </tr>\n","    <tr>\n","      <th>Sector_naics_45</th>\n","      <th>1</th>\n","      <td>-0.334850</td>\n","      <td>1.938566</td>\n","    </tr>\n","    <tr>\n","      <th>Sector_naics_48</th>\n","      <th>1</th>\n","      <td>-0.351157</td>\n","      <td>0.588435</td>\n","    </tr>\n","    <tr>\n","      <th>Sector_naics_49</th>\n","      <th>1</th>\n","      <td>-0.193406</td>\n","      <td>1.002416</td>\n","    </tr>\n","    <tr>\n","      <th>Sector_naics_51</th>\n","      <th>1</th>\n","      <td>-0.191845</td>\n","      <td>0.982884</td>\n","    </tr>\n","    <tr>\n","      <th>Sector_naics_52</th>\n","      <th>1</th>\n","      <td>-0.346630</td>\n","      <td>0.494335</td>\n","    </tr>\n","    <tr>\n","      <th>Sector_naics_53</th>\n","      <th>1</th>\n","      <td>-0.330990</td>\n","      <td>0.515588</td>\n","    </tr>\n","    <tr>\n","      <th>Sector_naics_54</th>\n","      <th>1</th>\n","      <td>-0.252519</td>\n","      <td>0.670019</td>\n","    </tr>\n","    <tr>\n","      <th>Sector_naics_56</th>\n","      <th>1</th>\n","      <td>-0.318486</td>\n","      <td>0.642288</td>\n","    </tr>\n","    <tr>\n","      <th>Sector_naics_61</th>\n","      <th>1</th>\n","      <td>-0.264770</td>\n","      <td>0.354230</td>\n","    </tr>\n","    <tr>\n","      <th>Sector_naics_62</th>\n","      <th>1</th>\n","      <td>-0.229430</td>\n","      <td>1.072065</td>\n","    </tr>\n","    <tr>\n","      <th>Sector_naics_71</th>\n","      <th>1</th>\n","      <td>-0.468323</td>\n","      <td>1.456940</td>\n","    </tr>\n","    <tr>\n","      <th>Sector_naics_72</th>\n","      <th>1</th>\n","      <td>-0.448787</td>\n","      <td>1.185426</td>\n","    </tr>\n","    <tr>\n","      <th>Sector_naics_81</th>\n","      <th>1</th>\n","      <td>-0.480759</td>\n","      <td>0.742218</td>\n","    </tr>\n","    <tr>\n","      <th>Sector_naics_99</th>\n","      <th>1</th>\n","      <td>-0.312724</td>\n","      <td>0.222852</td>\n","    </tr>\n","  </tbody>\n","</table>\n","</div>"],"text/plain":["                       RetEarly2020  RetLate2020\n","variable        value                           \n","Sector_naics_11 1         -0.158294     0.123019\n","Sector_naics_21 1         -0.597364     1.231517\n","Sector_naics_22 1         -0.155907     0.191994\n","Sector_naics_23 1         -0.371588     1.090058\n","Sector_naics_31 1         -0.281907     0.775376\n","Sector_naics_32 1         -0.182609     0.760634\n","Sector_naics_33 1         -0.293057     1.022835\n","Sector_naics_42 1         -0.320072     0.755734\n","Sector_naics_44 1         -0.388449     1.527506\n","Sector_naics_45 1         -0.334850     1.938566\n","Sector_naics_48 1         -0.351157     0.588435\n","Sector_naics_49 1         -0.193406     1.002416\n","Sector_naics_51 1         -0.191845     0.982884\n","Sector_naics_52 1         -0.346630     0.494335\n","Sector_naics_53 1         -0.330990     0.515588\n","Sector_naics_54 1         -0.252519     0.670019\n","Sector_naics_56 1         -0.318486     0.642288\n","Sector_naics_61 1         -0.264770     0.354230\n","Sector_naics_62 1         -0.229430     1.072065\n","Sector_naics_71 1         -0.468323     1.456940\n","Sector_naics_72 1         -0.448787     1.185426\n","Sector_naics_81 1         -0.480759     0.742218\n","Sector_naics_99 1         -0.312724     0.222852"]},"metadata":{"tags":[]},"execution_count":243}]},{"cell_type":"code","metadata":{"colab":{"base_uri":"https://localhost:8080/","height":387},"id":"T6PLWkgwByrm","executionInfo":{"status":"ok","timestamp":1625713113722,"user_tz":240,"elapsed":461,"user":{"displayName":"Tzu-Hua Huang","photoUrl":"https://lh3.googleusercontent.com/a-/AOh14GgbXOdZWwXGVTDuMHbeSme5C5oKAeF6I6ojY6qc=s64","userId":"01458545184256376706"}},"outputId":"9151713c-ec79-4cf7-f6c5-c8f0f25e797b"},"source":["# barplot\n","df_naics_AvgRet.query('value == 1').plot(kind='bar')"],"execution_count":244,"outputs":[{"output_type":"execute_result","data":{"text/plain":["<matplotlib.axes._subplots.AxesSubplot at 0x7f81b6fc9e90>"]},"metadata":{"tags":[]},"execution_count":244},{"output_type":"display_data","data":{"image/png":"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\n","text/plain":["<Figure size 432x288 with 1 Axes>"]},"metadata":{"tags":[],"needs_background":"light"}}]},{"cell_type":"markdown","metadata":{"id":"itdO_DXtqTlw"},"source":["**RetEarly2020**\n","\n","- Among all the industries, the Utilities sector (`Sector_naics_45`) has the highest return rate in 2020, indicating the magnitude telecommunication sector impacted by the pandemic is the least. However, the negative change of return suggests the Utilities sector faced a loss in early 2020. \n","- Compared to the other industries, the Mining, Quarrying, and Oil and Gas Extraction sector (`Sector_naics_21`) has the lowest returns in 2020. During the pandemic, people were confined to their homes and restricted to travel, resulting in a declined oil demand and a decrease in returns, which made the energy sector the sector impacted by the pandemic the worst.\n","\n","**RetLate2020**\n","- For late 2020, the Retail Trade sector (`Sector_naics_45`) has the highest return compared to the other industries. The pandemic caused a consumption shock due to the lockdown and forced shutdown across several industries. However, during recovering from the pandemic in 2020, people may have been accumulating more savings while forced to cut back on spending from lockdowns and travel restrictions. Thus, in late 2020, the Retail Trade sector may have the highest bounce in return.\n","- In contrast, the Agriculture, Forestry, Fishing and Hunting sector (`Sector_naics_21`) has the lowest return growth among all the industries."]},{"cell_type":"markdown","metadata":{"id":"Xv5We6KmqTlw"},"source":["#### ii. Fixed effect regressions\n","Run fixed effect regressions (using industry fixed effects) to explain both *ReturnEarly2020* and *ReturnLate2020:*\n","\n","1. You may wish to review the lecture notes from Lecture #3 and use the Colab Notebook: Fixed_Effect_Regression_for_BA870_Lecture_3.ipynb that is posted on QTools in Lecture #3 folder.\n","2. What is the explanatory power of these regressions using different types of industry variables. Is this high or low?Explain.\n","\n","Combine the industry fixed effects with other explanatory variables (see Assignment #3 and/or the financial ratios described in OPTION (3) below).\n","   - Run combined regressions with both industry and other “X” variables."]},{"cell_type":"markdown","metadata":{"id":"4ajQcv2yqTlw"},"source":["**RetEarly2020**"]},{"cell_type":"code","metadata":{"colab":{"base_uri":"https://localhost:8080/"},"id":"uOGGwnpfqTlw","executionInfo":{"status":"ok","timestamp":1625713113724,"user_tz":240,"elapsed":54,"user":{"displayName":"Tzu-Hua Huang","photoUrl":"https://lh3.googleusercontent.com/a-/AOh14GgbXOdZWwXGVTDuMHbeSme5C5oKAeF6I6ojY6qc=s64","userId":"01458545184256376706"}},"outputId":"2662f757-a9d2-4445-ffcb-4089466e062a"},"source":["# define x as a subset of original dataframe\n","x1_naics = df_naics.drop(columns=['TICKER', 'RetEarly2020', 'RetLate2020', 'naics'])\n","# define y as a series (RetEarly2020)\n","y1_naics = df_naics['RetEarly2020']\n","\n","# run fixed effect regression\n","# pass x as a dataframe, while pass y as a series\n","model1_naics = sm.OLS(y1_naics, x1_naics).fit()\n","\n","# print out regression summary statistics\n","print(model1_naics.summary())"],"execution_count":245,"outputs":[{"output_type":"stream","text":["                            OLS Regression Results                            \n","==============================================================================\n","Dep. Variable:           RetEarly2020   R-squared:                       0.083\n","Model:                            OLS   Adj. R-squared:                  0.075\n","Method:                 Least Squares   F-statistic:                     10.88\n","Date:                Thu, 08 Jul 2021   Prob (F-statistic):           1.42e-36\n","Time:                        02:58:33   Log-Likelihood:                -572.06\n","No. Observations:                2665   AIC:                             1190.\n","Df Residuals:                    2642   BIC:                             1326.\n","Df Model:                          22                                         \n","Covariance Type:            nonrobust                                         \n","===================================================================================\n","                      coef    std err          t      P>|t|      [0.025      0.975]\n","-----------------------------------------------------------------------------------\n","Sector_naics_11    -0.1583      0.123     -1.287      0.198      -0.399       0.083\n","Sector_naics_21    -0.5974      0.034    -17.403      0.000      -0.665      -0.530\n","Sector_naics_22    -0.1559      0.036     -4.331      0.000      -0.227      -0.085\n","Sector_naics_23    -0.3716      0.049     -7.605      0.000      -0.467      -0.276\n","Sector_naics_31    -0.2819      0.033     -8.578      0.000      -0.346      -0.217\n","Sector_naics_32    -0.1826      0.014    -12.688      0.000      -0.211      -0.154\n","Sector_naics_33    -0.2931      0.013    -21.886      0.000      -0.319      -0.267\n","Sector_naics_42    -0.3201      0.038     -8.501      0.000      -0.394      -0.246\n","Sector_naics_44    -0.3884      0.036    -10.867      0.000      -0.459      -0.318\n","Sector_naics_45    -0.3348      0.053     -6.289      0.000      -0.439      -0.230\n","Sector_naics_48    -0.3512      0.037     -9.471      0.000      -0.424      -0.278\n","Sector_naics_49    -0.1934      0.213     -0.908      0.364      -0.611       0.224\n","Sector_naics_51    -0.1918      0.018    -10.446      0.000      -0.228      -0.156\n","Sector_naics_52    -0.3466      0.013    -26.014      0.000      -0.373      -0.321\n","Sector_naics_53    -0.3310      0.022    -15.384      0.000      -0.373      -0.289\n","Sector_naics_54    -0.2525      0.036     -7.064      0.000      -0.323      -0.182\n","Sector_naics_56    -0.3185      0.044     -7.171      0.000      -0.406      -0.231\n","Sector_naics_61    -0.2648      0.087     -3.045      0.002      -0.435      -0.094\n","Sector_naics_62    -0.2294      0.050     -4.633      0.000      -0.327      -0.132\n","Sector_naics_71    -0.4683      0.087     -5.386      0.000      -0.639      -0.298\n","Sector_naics_72    -0.4488      0.043    -10.323      0.000      -0.534      -0.364\n","Sector_naics_81    -0.4808      0.114     -4.223      0.000      -0.704      -0.258\n","Sector_naics_99    -0.3127      0.213     -1.468      0.142      -0.730       0.105\n","==============================================================================\n","Omnibus:                     4130.784   Durbin-Watson:                   1.863\n","Prob(Omnibus):                  0.000   Jarque-Bera (JB):          4143167.282\n","Skew:                           9.309   Prob(JB):                         0.00\n","Kurtosis:                     195.263   Cond. No.                         16.0\n","==============================================================================\n","\n","Warnings:\n","[1] Standard Errors assume that the covariance matrix of the errors is correctly specified.\n"],"name":"stdout"}]},{"cell_type":"markdown","metadata":{"id":"KrN5P7mcqTlw"},"source":["**R-squared & Adj. R-squared**\n","\n","- `R-squared`: 0.083\n","\n","Around 8% of the variation in `RetEarly2020` is explained by the industries that the companies belong to. We can infer that returns in early 2020 are described by its business model of the type of business that operates. So 8% of the variation which varies across all these companies is explained by the sector itself.\n","\n","- `Adjusted R-squared`: 0.075\n","\n","However, the Adjusted R-squared is less than the R-squared, which is 0.075. Since Adjusted R-squared considers different independent variables against the model, a decrease inA djusted R-squared indicates the additional variables doen't improve the model more than would be expected by chance.\n","\n","**t-stats**\n","\n","Among all the variables, except `Sector_naics_49` and `Sector_naics_99`, all the sector variables have large t-values, indicating the most significant difference compared to the other models. Their p-values are also far less than 0.05. We can reject the null hypothesis and conclude that the data provide enough evidence to reject the null hypothesis, suggesting that each variable is statistically significant. Data favor the hypothesis that there is a non-zero correlation between almost all the variables and `RetEarly2020`."]},{"cell_type":"markdown","metadata":{"id":"03EgihXuqTlx"},"source":["**RetLate2020**"]},{"cell_type":"code","metadata":{"colab":{"base_uri":"https://localhost:8080/"},"id":"sI-TCz2sqTlx","executionInfo":{"status":"ok","timestamp":1625713113736,"user_tz":240,"elapsed":62,"user":{"displayName":"Tzu-Hua Huang","photoUrl":"https://lh3.googleusercontent.com/a-/AOh14GgbXOdZWwXGVTDuMHbeSme5C5oKAeF6I6ojY6qc=s64","userId":"01458545184256376706"}},"outputId":"7318d124-a228-4efb-a566-b72467975dc2"},"source":["# define x as a subset of original dataframe\n","x2_naics = df_naics.drop(columns=['TICKER', 'RetEarly2020', 'RetLate2020', 'naics'])\n","# define y as a series (RetLate2020)\n","y2_naics = df_naics['RetLate2020']\n","\n","# run fixed effect regression\n","# pass x as a dataframe, while pass y as a series\n","model2_naics = sm.OLS(y2_naics, x2_naics).fit()\n","\n","# print out regression summary statistics\n","print(model2_naics.summary())"],"execution_count":246,"outputs":[{"output_type":"stream","text":["                            OLS Regression Results                            \n","==============================================================================\n","Dep. Variable:            RetLate2020   R-squared:                       0.080\n","Model:                            OLS   Adj. R-squared:                  0.073\n","Method:                 Least Squares   F-statistic:                     10.50\n","Date:                Thu, 08 Jul 2021   Prob (F-statistic):           4.51e-35\n","Time:                        02:58:33   Log-Likelihood:                -3859.9\n","No. Observations:                2665   AIC:                             7766.\n","Df Residuals:                    2642   BIC:                             7901.\n","Df Model:                          22                                         \n","Covariance Type:            nonrobust                                         \n","===================================================================================\n","                      coef    std err          t      P>|t|      [0.025      0.975]\n","-----------------------------------------------------------------------------------\n","Sector_naics_11     0.1230      0.422      0.291      0.771      -0.705       0.951\n","Sector_naics_21     1.2315      0.118     10.448      0.000       1.000       1.463\n","Sector_naics_22     0.1920      0.124      1.553      0.121      -0.050       0.434\n","Sector_naics_23     1.0901      0.168      6.496      0.000       0.761       1.419\n","Sector_naics_31     0.7754      0.113      6.870      0.000       0.554       0.997\n","Sector_naics_32     0.7606      0.049     15.390      0.000       0.664       0.858\n","Sector_naics_33     1.0228      0.046     22.244      0.000       0.933       1.113\n","Sector_naics_42     0.7557      0.129      5.845      0.000       0.502       1.009\n","Sector_naics_44     1.5275      0.123     12.444      0.000       1.287       1.768\n","Sector_naics_45     1.9386      0.183     10.602      0.000       1.580       2.297\n","Sector_naics_48     0.5884      0.127      4.622      0.000       0.339       0.838\n","Sector_naics_49     1.0024      0.731      1.371      0.171      -0.432       2.437\n","Sector_naics_51     0.9829      0.063     15.585      0.000       0.859       1.107\n","Sector_naics_52     0.4943      0.046     10.804      0.000       0.405       0.584\n","Sector_naics_53     0.5156      0.074      6.979      0.000       0.371       0.660\n","Sector_naics_54     0.6700      0.123      5.458      0.000       0.429       0.911\n","Sector_naics_56     0.6423      0.153      4.212      0.000       0.343       0.941\n","Sector_naics_61     0.3542      0.299      1.186      0.236      -0.231       0.940\n","Sector_naics_62     1.0721      0.170      6.305      0.000       0.739       1.405\n","Sector_naics_71     1.4569      0.299      4.879      0.000       0.871       2.042\n","Sector_naics_72     1.1854      0.149      7.940      0.000       0.893       1.478\n","Sector_naics_81     0.7422      0.391      1.899      0.058      -0.024       1.509\n","Sector_naics_99     0.2229      0.731      0.305      0.761      -1.211       1.657\n","==============================================================================\n","Omnibus:                     2937.219   Durbin-Watson:                   1.952\n","Prob(Omnibus):                  0.000   Jarque-Bera (JB):           390055.460\n","Skew:                           5.363   Prob(JB):                         0.00\n","Kurtosis:                      61.289   Cond. No.                         16.0\n","==============================================================================\n","\n","Warnings:\n","[1] Standard Errors assume that the covariance matrix of the errors is correctly specified.\n"],"name":"stdout"}]},{"cell_type":"markdown","metadata":{"id":"-3ufsoj0qTlx"},"source":["**R-squared & Adj. R-squared**\n","\n","- `R-squared`: 0.080\n","\n","  Around 8% of the variation in `RetLate2020` is explained by the industries that the companies belong to. We can infer that returns in late 2020 are described by its business model of the type of business that operates. So 8% of the variation which varies across all these companies is explained by the sector itself.\n","\n","- `Adjusted R-squared`: 0.073\n","\n","  However, the Adjusted R-squared decreases compared to the R-squared, which is 0.073, indicating that the increase in the variables in the model doesn't enhance the explanatory power of the model. \n","\n","**t-stats**\n","\n","Compared to the regression model for `RetEarly2020`, there are more variables have large p-values. In general, most of the variables have large t-values, and their p-values are also far less than 0.05. We can reject the null hypothesis and conclude that the data provide enough evidence to reject the null hypothesis, suggesting that each variable is statistically significant. Data favor the hypothesis that there is a non-zero correlation between almost all the variables and `RetLate2020`. "]}]}