Summer_camp.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {
    "pycharm": {
     "name": "#%% md\n"
    }
   },
   "source": [
    "<h2 style = 'text-align: center'>\n",
    "Divar Data Analysis Summer Camp Entrance Task\n",
    "</h2>\n",
    "<h3 style = 'text-align: center'> By : Fatemeh Shiri </h3>\n",
    "<h4 style = 'text-align: center'> Gmail Address : rominashiri@gmail.com </h4>\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 46,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "import pandas as pd\n",
    "import numpy as np\n",
    "import matplotlib.pyplot as plt\n",
    "import seaborn as sns"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 47,
   "metadata": {
    "pycharm": {
     "name": "#%%\n"
    }
   },
   "outputs": [],
   "source": [
    "df = pd.read_excel('Summer Camp Task Data.xlsx')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 48,
   "metadata": {
    "pycharm": {
     "name": "#%%\n"
    }
   },
   "outputs": [],
   "source": [
    "df['created_at'] = pd.to_datetime(df['created_at'],unit='ms')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {
    "pycharm": {
     "name": "#%%\n"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "action                         0\n",
       "created_at                     0\n",
       "source_event_id                0\n",
       "device_id                   1487\n",
       "post_page_offset           75796\n",
       "tokens                     75796\n",
       "post_index_in_post_list    35287\n",
       "post_token                 35288\n",
       "dtype: int64"
      ]
     },
     "execution_count": 5,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df.isnull().sum()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "pycharm": {
     "name": "#%% md\n"
    }
   },
   "source": [
    "----"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {
    "pycharm": {
     "name": "#%%\n"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(35287, 75796)"
      ]
     },
     "execution_count": 6,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "len(df[df['action']=='load_post_page']) , len(df[df['action']=='click_post'])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 83,
   "metadata": {
    "pycharm": {
     "name": "#%%\n"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.axes._subplots.AxesSubplot at 0x1fbf6891850>"
      ]
     },
     "execution_count": 83,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 720x432 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "sns.set_style('darkgrid')\n",
    "plt.figure(figsize=(10,6))\n",
    "df['post_page_offset'].hist(bins=25)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "pycharm": {
     "name": "#%% md\n"
    }
   },
   "source": [
    "---"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {
    "pycharm": {
     "name": "#%%\n"
    }
   },
   "outputs": [],
   "source": [
    "Times_seies_df = df.set_index('created_at',drop=True)\n",
    "Times_seies_df['action'] = Times_seies_df['action'].apply(lambda x:1 if x=='load_post_page' else 0)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 49,
   "metadata": {
    "pycharm": {
     "name": "#%%\n"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.axes._subplots.AxesSubplot at 0x22be21f6850>"
      ]
     },
     "execution_count": 49,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "text/plain": [
       "<Figure size 720x432 with 0 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 720x432 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "sns.set_style('darkgrid')\n",
    "plt.figure(figsize=(10,6))\n",
    "Times_seies_df.resample('H').count().plot(figsize=(10,6))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "---"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "pycharm": {
     "name": "#%% md\n"
    }
   },
   "source": [
    "<h2 style='direction:rtl;text-align:right;font-family:B Lotus, B Nazanin, Tahoma'> سوال #1\n",
    "</h2>\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "pycharm": {
     "name": "#%% md\n"
    }
   },
   "source": [
    "<h3 style='direction:rtl;text-align:right;font-family:B Lotus, B Nazanin, Tahoma;font-size:20px'>\n",
    "    <b>\n",
    "خطاهای مربوط به داده ها:\n",
    "    </b>\n",
    "</h3>\n",
    "\n",
    "<ul style='direction:rtl;text-align:right;font-family:B Lotus, B Nazanin, Tahoma;font-size:18px'>\n",
    "    <li style=direction:rtl>برای تعدادی از داده ها صرفا اکشن کلیک،بدون اکشن لود ثبت شده است. </li>\n",
    "    <br>\n",
    "    <li style=direction:rtl>تعداد زیادی device_id  ثبت نشده داریم که برای بررسی یکتا بودن source_event  مشکل ایجاد میکند. </li>\n",
    "    <br>\n",
    "    <li style=direction:rtl>برای یک source_event  یکسان تعداد زیادیload_post_page داریم که نامتناسب و عجیب به نظر میرسد. </li>\n",
    "    <br>\n",
    "\n",
    "</ul>"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "---"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "pycharm": {
     "name": "#%% md\n"
    }
   },
   "source": [
    "<h2 style='direction:rtl;text-align:right;font-family:B Lotus, B Nazanin, Tahoma'> سوال #2\n",
    "</h2>\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 40,
   "metadata": {
    "pycharm": {
     "name": "#%%\n"
    }
   },
   "outputs": [],
   "source": [
    "def get_number_of_results(result_token):\n",
    "    if result_token != np.nan:\n",
    "        results = str(result_token).split(',')\n",
    "        return len(results)\n",
    "    else:\n",
    "        return np.nan\n",
    "\n",
    "df['number_of_results'] = df['tokens'].apply(lambda x:get_number_of_results(x))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "<h4 style='direction:rtl;text-align:right;font-family:B Lotus, B Nazanin, Tahoma'> \n",
    "    متریک dark query percent \n",
    "</h4>"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {
    "pycharm": {
     "name": "#%%\n"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "3.3723467565959133"
      ]
     },
     "execution_count": 15,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "dark_query_percent = len(df[(df['action']=='load_post_page')&(df['number_of_results']<10)&(df['post_page_offset']==0)])/len(df[(df['action']=='load_post_page')])*100\n",
    "dark_query_percent"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "pycharm": {
     "name": "#%% md\n"
    }
   },
   "source": [
    "<h4 style='direction:rtl;text-align:right;font-family:B Lotus, B Nazanin, Tahoma'> \n",
    "    متریک query bounce rate \n",
    "</h4>"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {
    "pycharm": {
     "name": "#%%\n"
    }
   },
   "outputs": [],
   "source": [
    "source_df = df.groupby(by=['source_event_id','action']).count()['number_of_results']\n",
    "clicked = source_df.reset_index().groupby('source_event_id').count()['action'].value_counts()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 193,
   "metadata": {
    "pycharm": {
     "name": "#%%\n"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "23.779295491257404"
      ]
     },
     "execution_count": 193,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "query_bounce_rate = clicked[1]/len(df[df['action']=='load_post_page'])*100\n",
    "query_bounce_rate"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "---"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "<h2 style='direction:rtl;text-align:right;font-family:B Lotus, B Nazanin, Tahoma'> سوال #3\n",
    "</h2>\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "metadata": {
    "pycharm": {
     "name": "#%%\n"
    }
   },
   "outputs": [],
   "source": [
    "given_query = 'ffef6048-9db0-4244-9750-3b530552f0ec'"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "<h4 style='direction:rtl;text-align:right;font-family:B Lotus, B Nazanin, Tahoma'> \n",
    "<strong>\n",
    "    متریک درصد آگهی های کلیک شده نسبت به آگهی های لود شده \n",
    "</strong>\n",
    "</h4>\n",
    "\n",
    "<br>\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {
    "pycharm": {
     "name": "#%%\n"
    }
   },
   "outputs": [],
   "source": [
    "def get_click_percent(given_query):\n",
    "    clicked_loaded_df = df[df['source_event_id']== given_query]\n",
    "    if 'load_post_page' in list(clicked_loaded_df['action']):\n",
    "        loaded = clicked_loaded_df[(clicked_loaded_df['action']=='load_post_page')]\n",
    "        if 'click_post' in list(clicked_loaded_df['action']):\n",
    "            clicked = clicked_loaded_df[clicked_loaded_df['action']=='click_post']\n",
    "        else:\n",
    "            clicked = np.nan\n",
    "    else:\n",
    "        loaded = np.nan\n",
    "\n",
    "    if clicked & loaded != np.nan:\n",
    "        return len(clicked)/(loaded['number_of_results']*loaded['post_page_offset'])*100\n",
    "    else:\n",
    "        return 'no load or click'"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "pycharm": {
     "name": "#%%\n"
    }
   },
   "source": [
    "<h4 style='direction:rtl;text-align:right;font-family:B Lotus, B Nazanin, Tahoma'> \n",
    "<strong>\n",
    "    متریک رتبه اولین کلیک کاربر \n",
    "</strong>\n",
    "</h4>\n",
    "\n",
    "<br>\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 42,
   "metadata": {
    "pycharm": {
     "name": "#%%\n"
    }
   },
   "outputs": [],
   "source": [
    "def get_first_click(given_query):\n",
    "    clicked_df = df[df['source_event_id']== given_query]\n",
    "    clicked_df = clicked_df.sort_values('created_at')\n",
    "    if 'click_post' in list(clicked_df['action']):\n",
    "        return clicked_df[clicked_df['action']=='click_post'].iloc[0]['post_index_in_post_list']\n",
    "    else:\n",
    "        return np.nan"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "pycharm": {
     "name": "#%%\n"
    }
   },
   "source": [
    "<h4 style='direction:rtl;text-align:right;font-family:B Lotus, B Nazanin, Tahoma'> \n",
    "<strong>\n",
    "    متریک میانگین فاصله بین رتبه کلیک های کاربر\n",
    "</strong>\n",
    "</h4>\n",
    "\n",
    "<br>\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {
    "pycharm": {
     "name": "#%%\n"
    }
   },
   "outputs": [],
   "source": [
    "def get_clicks_Avg_mark(given_query):\n",
    "    clicked_df = df[df['source_event_id']== given_query]\n",
    "    clicked_df = clicked_df.sort_values('created_at')\n",
    "    marks = []\n",
    "    if 'click_post' in list(clicked_df['action']):\n",
    "        marks.append(clicked_df[clicked_df['action']=='click_post']['post_index_in_post_list'])\n",
    "        mark = sum(sum(marks))/len(marks)\n",
    "        return mark\n",
    "    else:\n",
    "        return 'no click'"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "pycharm": {
     "name": "#%%\n"
    }
   },
   "source": [
    "<h4 style='direction:rtl;text-align:right;font-family:B Lotus, B Nazanin, Tahoma'> \n",
    "<strong>\n",
    "اینکه آیا روی یکی از ۳ نتیجه اول کوئری کلیک شده یا نه</strong>\n",
    "</h4>\n",
    "\n",
    "<br>\n",
    "\n",
    "\n",
    "\n",
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {
    "pycharm": {
     "name": "#%%\n"
    }
   },
   "outputs": [],
   "source": [
    "def check_first_3_posts(qiven_query):\n",
    "    clicked_df = df[df['source_event_id']== given_query]\n",
    "    posts = []\n",
    "    check = False\n",
    "    if 'click_post' in list(clicked_df['action']):\n",
    "        posts = (clicked_df[clicked_df['action']=='click_post'].iloc[:]['post_index_in_post_list'])\n",
    "        for i in posts:\n",
    "            if i < 3 :\n",
    "                check = True\n",
    "\n",
    "        if check:\n",
    "            return True\n",
    "    else:\n",
    "        return False"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 43,
   "metadata": {
    "pycharm": {
     "name": "#%%\n"
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Wall time: 0 ns\n"
     ]
    }
   ],
   "source": [
    "%time\n",
    "df['First_click'] = df['source_event_id'].apply(lambda x:get_first_click(str(x)))\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 45,
   "metadata": {
    "pycharm": {
     "name": "#%%\n"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "56.181102362204726"
      ]
     },
     "execution_count": 45,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df['First_click'].mean()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "pycharm": {
     "name": "#%% md\n"
    }
   },
   "source": [
    "<h3 style='direction:rtl;text-align:right;font-family:B Lotus, B Nazanin, Tahoma'>\n",
    "به نظر من\n",
    "    <i >\n",
    "    متریک رتبه اولین کلیک کاربر\n",
    "    </i>\n",
    "    متریک مناسب تری است.\n",
    "<br>\n",
    "در واقع برای انتخاب این متریک نقاط ضعف و قوت تمام متریک ها را کنار هم قرار دادم و به این نتیجه رسیدم ولی به طور کلی واضح است که\n",
    "همه این متریک ها کنار هم مانند مدل سوال 4 و یا وابسته به کاربرد و مسئله مورد نظر ما بهتر و بهینه تر قابل استفاده هستند.\n",
    "    <br>\n",
    "    در ادامه نکات مربوط به هر متریک را تیتر وار بیان میکنم:\n",
    "\n",
    "</h3>\n",
    "<ul style='direction:rtl;text-align:right;font-family:B Lotus, B Nazanin, Tahoma;font-size:18px'>\n",
    "    <br>\n",
    "<li>\n",
    "متریک درصد آگهی کلیک شده نسبت به آگهی  لود شده نمیتواند به ما خروجی ای ارائه دهد که بر اساس آن عملکرد را بهبود ببخشیم.\n",
    "نتیجه گیری ساده از کم بودن این متریک میتواند این باشد که کاربر کوئری درست و مد نظر خود را جستجو نکرده است، که در حالت ساده رویکرد مناسبی برای بهبود آن پیشنهاد نمیدهد.\n",
    "یا خیلی بالا بودن این متریک نشان دهنده این است که کاربر روی تعداد زیادی اگهی کلیک کرده که ممکن است رفتار این کاربر صرفا برای دیدن جزئیات اگهی ها بوده باشد و در نتیجه قابل استناد نیست.\n",
    "</li>\n",
    "    <br>\n",
    "<li>\n",
    "متریک میانگین فاصله بین رتبه کلیک های کاربر ممکن است در مواجهه با رتبه کلیک های unbalanced  ضعیف عمل کند و دید نامناسبی روی کلیک های کاربر میدهد.\n",
    "همچنین ممکن است برخی یوزر ها روی تعداد زیادی اگهی کلیک کنند که این میانگین فاصله را به سمت عدد ثابتی میبرد.\n",
    "</li>\n",
    "    <br>\n",
    "<li>\n",
    "متریک کلیک روی 3 نتیجه اول هم بدلیل اینکه نمایش خروجی ها همیشه بر حسب زمان آخرین آگهی بوده عملا میتواند نشان دهد که این رویکرد نمایش آگهی در این موقعیت مناسب هست یا نه.\n",
    "</li>\n",
    "    <br>\n",
    "<li>\n",
    "از آنجایی که در زمان کلیک کردن کاربر عادی بعد از یک overlook  روی تمام اگهی ها، آگهی ای را انتخاب میکند که بیشترین میزان fit شدن روی نظر یا کوئری اش را دارد\n",
    "پس میتوان از اولین کلیک کاربر فهمید کدام نوع آگهی ها برای او جذابیت بیشتری دارند و به خواسته او نزدیک تر هستند و بر این اساس میتوان به میزان منلسب بودن عملکرد کوئری و نحوه نمایش آگهی های بعدی تا حدودی پی برد .البته در این متریک هم اگر کاربر روی تعداد زیادی آگهی کلیک کرده باشد تا صرفا جزئیات درون آگهی ها را ببیند نامناسب عمل میکند، ولی قابلیت بهبود بیشتر و بهتری را ارائه میدهد.\n",
    "</li>\n",
    "</ul>\n",
    "\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "---"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "pycharm": {
     "name": "#%% md\n"
    }
   },
   "source": [
    "<h2 style='direction:rtl;text-align:right;font-family:B Lotus, B Nazanin, Tahoma'> سوال #4\n",
    "</h2>"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "pycharm": {
     "name": "#%% md\n"
    }
   },
   "source": [

   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.8.3"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 1
}