Housing_Price_Predictor_AWS.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Matplotlib is building the font cache; this may take a moment.\n"
     ]
    }
   ],
   "source": [
    "import pandas as pd\n",
    "import numpy as np\n",
    "import matplotlib.pyplot as plt\n",
    "from sklearn.metrics import mean_squared_error, mean_absolute_error, median_absolute_error, root_mean_squared_error\n",
    "from sklearn.model_selection import train_test_split\n",
    "from sklearn.metrics import accuracy_score\n",
    "from sagemaker.image_uris import retrieve\n",
    "from sagemaker.predictor import Predictor\n",
    "from sagemaker import image_uris\n",
    "import boto3\n",
    "import io\n",
    "import os\n",
    "import sagemaker"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "nelson = pd.read_csv('datasets/regional/Nelson.csv',index_col=False)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [
    {
     "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>Nelson</th>\n",
       "      <th>lagged</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>date</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>1992-01-31</th>\n",
       "      <td>115000.0</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1992-02-29</th>\n",
       "      <td>108000.0</td>\n",
       "      <td>115000.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1992-03-31</th>\n",
       "      <td>115000.0</td>\n",
       "      <td>108000.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1992-04-30</th>\n",
       "      <td>116000.0</td>\n",
       "      <td>115000.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1992-05-31</th>\n",
       "      <td>115000.0</td>\n",
       "      <td>116000.0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "              Nelson    lagged\n",
       "date                          \n",
       "1992-01-31  115000.0       NaN\n",
       "1992-02-29  108000.0  115000.0\n",
       "1992-03-31  115000.0  108000.0\n",
       "1992-04-30  116000.0  115000.0\n",
       "1992-05-31  115000.0  116000.0"
      ]
     },
     "execution_count": 4,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "nelson.index = nelson.date\n",
    "nelson.index = pd.to_datetime(nelson.index)\n",
    "nelson = nelson[['Nelson', 'lagged']]\n",
    "nelson.head(5)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "---"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## AWS"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Using DeepAR"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "The data will be converted first to quarterly data as DeepAR uses multiple data points per timestep i.e, if using weekly timesteps, daily data per timestep is required. In this case we will use three months of data (3 values) per timestep (quarter)."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 70,
   "metadata": {},
   "outputs": [],
   "source": [
    "def months_to_quarters(df):\n",
    "    \n",
    "    df = df[:-2][['Nelson']]\n",
    "    df.rename(columns={'Nelson':'First_Month'}, inplace=True)\n",
    "    df['Second_Month'] = None\n",
    "    df['Third_Month'] = None\n",
    "    for i in range(2, len(df), 3):\n",
    "        s_month = df.loc[df.index[i-1],\"First_Month\"]\n",
    "        f_month = df.loc[df.index[i-2],\"First_Month\"]\n",
    "        df.loc[df.index[i], \"Third_Month\"] = df.loc[df.index[i], \"First_Month\"]\n",
    "        df.loc[df.index[i], \"Second_Month\"] = s_month\n",
    "        df.loc[df.index[i], \"First_Month\"] = f_month\n",
    "    return df.dropna()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 71,
   "metadata": {},
   "outputs": [],
   "source": [
    "nelson_qrtly = months_to_quarters(nelson.copy())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 72,
   "metadata": {},
   "outputs": [
    {
     "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>First_Month</th>\n",
       "      <th>Second_Month</th>\n",
       "      <th>Third_Month</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>date</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>1992-03-31</th>\n",
       "      <td>115000.0</td>\n",
       "      <td>108000.0</td>\n",
       "      <td>115000.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1992-06-30</th>\n",
       "      <td>116000.0</td>\n",
       "      <td>115000.0</td>\n",
       "      <td>107000.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1992-09-30</th>\n",
       "      <td>118000.0</td>\n",
       "      <td>116000.0</td>\n",
       "      <td>120000.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1992-12-31</th>\n",
       "      <td>115000.0</td>\n",
       "      <td>122500.0</td>\n",
       "      <td>124000.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1993-03-31</th>\n",
       "      <td>120000.0</td>\n",
       "      <td>125000.0</td>\n",
       "      <td>126000.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2023-06-30</th>\n",
       "      <td>720000.0</td>\n",
       "      <td>770000.0</td>\n",
       "      <td>650000.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2023-09-30</th>\n",
       "      <td>710000.0</td>\n",
       "      <td>695000.0</td>\n",
       "      <td>680000.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2023-12-31</th>\n",
       "      <td>670500.0</td>\n",
       "      <td>650000.0</td>\n",
       "      <td>710000.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2024-03-31</th>\n",
       "      <td>675000.0</td>\n",
       "      <td>732500.0</td>\n",
       "      <td>722000.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2024-06-30</th>\n",
       "      <td>705000.0</td>\n",
       "      <td>670000.0</td>\n",
       "      <td>648000.0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>130 rows × 3 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "            First_Month Second_Month Third_Month\n",
       "date                                            \n",
       "1992-03-31     115000.0     108000.0    115000.0\n",
       "1992-06-30     116000.0     115000.0    107000.0\n",
       "1992-09-30     118000.0     116000.0    120000.0\n",
       "1992-12-31     115000.0     122500.0    124000.0\n",
       "1993-03-31     120000.0     125000.0    126000.0\n",
       "...                 ...          ...         ...\n",
       "2023-06-30     720000.0     770000.0    650000.0\n",
       "2023-09-30     710000.0     695000.0    680000.0\n",
       "2023-12-31     670500.0     650000.0    710000.0\n",
       "2024-03-31     675000.0     732500.0    722000.0\n",
       "2024-06-30     705000.0     670000.0    648000.0\n",
       "\n",
       "[130 rows x 3 columns]"
      ]
     },
     "execution_count": 72,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "nelson_qrtly"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 74,
   "metadata": {},
   "outputs": [],
   "source": [
    "train, test_and_validate = train_test_split(nelson_qrtly, \n",
    "                                            test_size=0.2, \n",
    "                                            random_state=42)\n",
    "\n",
    "train_test, validate = train_test_split(test_and_validate, \n",
    "                                  test_size=0.5, \n",
    "                                  random_state=42)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 75,
   "metadata": {},
   "outputs": [],
   "source": [
    "bucket='c125236a3097401l7344380t1w4575200910-sandboxbucket-6het1vfa8iyd'\n",
    "prefix='assignment3'\n",
    "train_file='train.json'\n",
    "test_file='test.json'\n",
    "validate_file='validate.json'\n",
    "whole_file='whole.json'\n",
    "s3_resource = boto3.Session().resource('s3')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 81,
   "metadata": {},
   "outputs": [],
   "source": [
    "import json\n",
    "import io\n",
    "\n",
    "def upload_s3_json(filename, folder, dataframe):\n",
    "    dates = [date.isoformat() for date in dataframe.index]\n",
    "    data = ''\n",
    "    for i in range(len(dataframe)):\n",
    "        temp = {}\n",
    "        temp[\"start\"] = dates[i]\n",
    "        temp[\"target\"] = [val for val in dataframe.iloc[i].values]\n",
    "        data += f'{json.dumps(temp)}\\n'\n",
    "    bytes_data = io.BytesIO(data.encode(\"utf-8\"))\n",
    "    s3_resource.Bucket(bucket).Object(os.path.join(prefix, folder, filename)).put(Body= bytes_data)\n",
    "\n",
    "upload_s3_json(train_file, 'train', train)\n",
    "upload_s3_json(test_file, 'test', train_test)\n",
    "upload_s3_json(validate_file, 'validate', validate)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 85,
   "metadata": {},
   "outputs": [],
   "source": [
    "role=sagemaker.get_execution_role()\n",
    "s3_output_location=\"s3://{}/{}/output/\".format(bucket,prefix)\n",
    "container = image_uris.retrieve(framework='forecasting-deepar', region=boto3.Session().region_name)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 86,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "'522234722520.dkr.ecr.us-east-1.amazonaws.com/forecasting-deepar:1'"
      ]
     },
     "execution_count": 86,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "container"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 91,
   "metadata": {},
   "outputs": [],
   "source": [
    "hyperparams={\n",
    "    \"time_freq\":\"3M\",\n",
    "    \"context_length\":\"1\",\n",
    "    \"prediction_length\":\"2\",\n",
    "    \"epochs\":\"100\"}"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 92,
   "metadata": {},
   "outputs": [],
   "source": [
    "deep_ar=sagemaker.estimator.Estimator(container,\n",
    "                                        role,\n",
    "                                        instance_count=1,\n",
    "                                        instance_type='ml.m5.xlarge',\n",
    "                                        output_path=s3_output_location,\n",
    "                                        hyperparameters=hyperparams,\n",
    "                                        sagemaker_session=sagemaker.Session())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 93,
   "metadata": {},
   "outputs": [],
   "source": [
    "train_channel = sagemaker.inputs.TrainingInput(\n",
    "    \"s3://{}/{}/train/\".format(bucket,prefix,train_file),\n",
    "    content_type='json')\n",
    "\n",
    "validate_channel = sagemaker.inputs.TrainingInput(\n",
    "    \"s3://{}/{}/validate/\".format(bucket,prefix,validate_file),\n",
    "    content_type='json')\n",
    "\n",
    "data_channels = {'train': train_channel, 'test': validate_channel}"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 95,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "INFO:sagemaker:Creating training-job with name: forecasting-deepar-2024-10-09-23-43-43-157\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "2024-10-09 23:43:45 Starting - Starting the training job..\n",
      "2024-10-09 23:44:02 Starting - Preparing the instances for training....\n",
      "2024-10-09 23:44:28 Downloading - Downloading input data.....\n",
      "2024-10-09 23:44:54 Downloading - Downloading the training image.............\n",
      "2024-10-09 23:46:05 Training - Training image download completed. Training in progress............\n",
      "2024-10-09 23:47:05 Uploading - Uploading generated training model\n",
      "2024-10-09 23:47:13 Completed - Training job completed\n"
     ]
    }
   ],
   "source": [
    "deep_ar.fit(inputs=data_channels, logs=False)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 97,
   "metadata": {},
   "outputs": [
    {
     "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>First_Month</th>\n",
       "      <th>Second_Month</th>\n",
       "      <th>Third_Month</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>date</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>2024-06-30</th>\n",
       "      <td>705000.0</td>\n",
       "      <td>670000.0</td>\n",
       "      <td>648000.0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "            First_Month Second_Month Third_Month\n",
       "date                                            \n",
       "2024-06-30     705000.0     670000.0    648000.0"
      ]
     },
     "execution_count": 97,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "nelson_qrtly[-1:]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 98,
   "metadata": {},
   "outputs": [],
   "source": [
    "sagemaker_session = sagemaker.Session()\n",
    "predictor = Predictor(endpoint_name='deep-ar-100', sagemaker_session=sagemaker_session)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 100,
   "metadata": {},
   "outputs": [],
   "source": [
    "data = {\n",
    "    \"instances\": [\n",
    "        {\n",
    "            \"start\": \"2024-06-30 00:00:00\",\n",
    "            \"target\": [705000.0, 670000.0, 648000.0],\n",
    "        }\n",
    "    ],\n",
    "    \"configuration\": {\n",
    "        \"num_samples\": 100,\n",
    "        \"output_types\": [\"mean\", \"quantiles\"],\n",
    "        \"quantiles\": [\"0.1\", \"0.5\", \"0.9\"]\n",
    "    }\n",
    "}\n",
    "\n",
    "payload = json.dumps(data)\n",
    "response = predictor.predict(payload, initial_args={\"ContentType\": \"application/json\"})"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 102,
   "metadata": {},
   "outputs": [],
   "source": [
    "result = json.loads(response.decode('utf-8'))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 107,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "{'predictions': [{'mean': [617007.875, 600444.5],\n",
       "   'quantiles': {'0.1': [556193.75, 543589.6875],\n",
       "    '0.5': [621433.5, 596463.375],\n",
       "    '0.9': [675815.125, 662208.0]}}]}"
      ]
     },
     "execution_count": 107,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "result"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 152,
   "metadata": {},
   "outputs": [],
   "source": [
    "means = [float(val) for val in result['predictions'][0]['mean']]\n",
    "q_01 = [float(val) for val in result['predictions'][0]['quantiles']['0.1']]\n",
    "q_05 = [float(val) for val in result['predictions'][0]['quantiles']['0.5']]\n",
    "q_09 = [float(val) for val in result['predictions'][0]['quantiles']['0.9']]\n",
    "index = [1,2]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 154,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Text(0.5, 0, 'Prediction')"
      ]
     },
     "execution_count": 154,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.plot(index, means)\n",
    "plt.plot(index, q_01)\n",
    "plt.plot(index, q_05)\n",
    "plt.plot(index, q_09)\n",
    "plt.title(\"0.1, 0.5, 0.9 Quantile and Mean Preictions\")\n",
    "plt.ylabel(\"Median Housing Price\")\n",
    "plt.xlabel(\"Prediction\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 133,
   "metadata": {},
   "outputs": [],
   "source": [
    "index = [val for val in nelson_qrtly.index]\n",
    "index.append(index[-1] + (pd.tseries.offsets.MonthEnd() * 3))\n",
    "index.append(index[-1] + (pd.tseries.offsets.MonthEnd() * 3))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 148,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Text(0.5, 1.0, 'Forecast Horizon for Median House price in NZ')"
      ]
     },
     "execution_count": 148,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.plot(index[100:-2], nelson_qrtly[100:].apply(np.mean, axis=1))\n",
    "plt.plot(index[-2:], q_01)\n",
    "plt.plot(index[-2:], q_05)\n",
    "plt.plot(index[-2:], q_09)\n",
    "plt.plot(index[-2:], means)\n",
    "plt.xlabel(\"Quarter\")\n",
    "plt.ylabel(\"Median House Price\")\n",
    "plt.title(\"Forecast Horizon for Median House price in NZ\")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Conclusion"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "It seems that the mean prediction is coming out far lower than expected, around the same value as the 0.5 quantile, whereas the 0.1 quantile is almost where one would expected the predicted value to lay. This model, using quarterly data, may simply be unable to capture the trend of the data."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "---"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "The data will now be converted to 2-Monthly intervals"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 159,
   "metadata": {},
   "outputs": [],
   "source": [
    "def months_to_2months(df):\n",
    "    \n",
    "    df = df[['Nelson']]\n",
    "    df.rename(columns={'Nelson':'First_Month'}, inplace=True)\n",
    "    df['Second_Month'] = None\n",
    "    for i in range(1, len(df), 2):\n",
    "        f_month = df.loc[df.index[i-1],\"First_Month\"]\n",
    "        df.loc[df.index[i], \"Second_Month\"] = df.loc[df.index[i], \"First_Month\"]\n",
    "        df.loc[df.index[i], \"First_Month\"] = f_month\n",
    "    return df.dropna()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 163,
   "metadata": {},
   "outputs": [],
   "source": [
    "nelson_bimonthly = months_to_2months(nelson.copy())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 164,
   "metadata": {},
   "outputs": [
    {
     "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>First_Month</th>\n",
       "      <th>Second_Month</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>date</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>1992-02-29</th>\n",
       "      <td>115000.0</td>\n",
       "      <td>108000.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1992-04-30</th>\n",
       "      <td>115000.0</td>\n",
       "      <td>116000.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1992-06-30</th>\n",
       "      <td>115000.0</td>\n",
       "      <td>107000.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1992-08-31</th>\n",
       "      <td>118000.0</td>\n",
       "      <td>116000.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1992-10-31</th>\n",
       "      <td>120000.0</td>\n",
       "      <td>115000.0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "            First_Month Second_Month\n",
       "date                                \n",
       "1992-02-29     115000.0     108000.0\n",
       "1992-04-30     115000.0     116000.0\n",
       "1992-06-30     115000.0     107000.0\n",
       "1992-08-31     118000.0     116000.0\n",
       "1992-10-31     120000.0     115000.0"
      ]
     },
     "execution_count": 164,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "nelson_bimonthly.head(5)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 165,
   "metadata": {},
   "outputs": [],
   "source": [
    "train, test = train_test_split(nelson_bimonthly, \n",
    "                                            test_size=0.2, \n",
    "                                            random_state=42)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 166,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "INFO:botocore.credentials:Found credentials from IAM Role: BaseNotebookInstanceEc2InstanceRole\n"
     ]
    }
   ],
   "source": [
    "bucket='c125236a3097401l7344380t1w4575200910-sandboxbucket-6het1vfa8iyd'\n",
    "prefix='assignment3'\n",
    "train_file='train_bi.json'\n",
    "test_file='test_bi.json'\n",
    "whole_file='whole_bi.json'\n",
    "s3_resource = boto3.Session().resource('s3')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 167,
   "metadata": {},
   "outputs": [],
   "source": [
    "upload_s3_json(train_file, 'train', train)\n",
    "upload_s3_json(test_file, 'test', test)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 168,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "INFO:sagemaker.image_uris:Same images used for training and inference. Defaulting to image scope: inference.\n",
      "INFO:sagemaker.image_uris:Defaulting to the only supported framework/algorithm version: 1.\n",
      "INFO:sagemaker.image_uris:Ignoring unnecessary instance type: None.\n"
     ]
    }
   ],
   "source": [
    "role=sagemaker.get_execution_role()\n",
    "s3_output_location=\"s3://{}/{}/output/\".format(bucket,prefix)\n",
    "container = image_uris.retrieve(framework='forecasting-deepar', region=boto3.Session().region_name)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 169,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "'522234722520.dkr.ecr.us-east-1.amazonaws.com/forecasting-deepar:1'"
      ]
     },
     "execution_count": 169,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "container"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 174,
   "metadata": {},
   "outputs": [],
   "source": [
    "hyperparams={\n",
    "    \"time_freq\":\"2M\",\n",
    "    \"context_length\":\"1\",\n",
    "    \"prediction_length\":\"1\",\n",
    "    \"epochs\":\"100\"}"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 175,
   "metadata": {},
   "outputs": [],
   "source": [
    "deep_ar=sagemaker.estimator.Estimator(container,\n",
    "                                        role,\n",
    "                                        instance_count=1,\n",
    "                                        instance_type='ml.m5.xlarge',\n",
    "                                        output_path=s3_output_location,\n",
    "                                        hyperparameters=hyperparams,\n",
    "                                        sagemaker_session=sagemaker.Session())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 176,
   "metadata": {},
   "outputs": [],
   "source": [
    "train_channel = sagemaker.inputs.TrainingInput(\n",
    "    \"s3://{}/{}/train/\".format(bucket,prefix,train_file),\n",
    "    content_type='json')\n",
    "\n",
    "test_channel = sagemaker.inputs.TrainingInput(\n",
    "    \"s3://{}/{}/validate/\".format(bucket,prefix,test_file),\n",
    "    content_type='json')\n",
    "\n",
    "data_channels = {'train': train_channel, 'test': test_channel}"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 177,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "INFO:sagemaker:Creating training-job with name: forecasting-deepar-2024-10-10-00-37-27-039\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "2024-10-10 00:37:28 Starting - Starting the training job..\n",
      "2024-10-10 00:37:44 Starting - Preparing the instances for training....\n",
      "2024-10-10 00:38:11 Downloading - Downloading input data.....\n",
      "2024-10-10 00:38:41 Downloading - Downloading the training image............\n",
      "2024-10-10 00:39:47 Training - Training image download completed. Training in progress..........\n",
      "2024-10-10 00:40:37 Uploading - Uploading generated training model..\n",
      "2024-10-10 00:40:50 Completed - Training job completed\n"
     ]
    }
   ],
   "source": [
    "deep_ar.fit(inputs=data_channels, logs=False)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 179,
   "metadata": {},
   "outputs": [
    {
     "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>First_Month</th>\n",
       "      <th>Second_Month</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>date</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>2023-12-31</th>\n",
       "      <td>650000.0</td>\n",
       "      <td>710000.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2024-02-29</th>\n",
       "      <td>675000.0</td>\n",
       "      <td>732500.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2024-04-30</th>\n",
       "      <td>722000.0</td>\n",
       "      <td>705000.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2024-06-30</th>\n",
       "      <td>670000.0</td>\n",
       "      <td>648000.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2024-08-31</th>\n",
       "      <td>657000.0</td>\n",
       "      <td>693000.0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "            First_Month Second_Month\n",
       "date                                \n",
       "2023-12-31     650000.0     710000.0\n",
       "2024-02-29     675000.0     732500.0\n",
       "2024-04-30     722000.0     705000.0\n",
       "2024-06-30     670000.0     648000.0\n",
       "2024-08-31     657000.0     693000.0"
      ]
     },
     "execution_count": 179,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "nelson_bimonthly.tail(5)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 180,
   "metadata": {},
   "outputs": [],
   "source": [
    "sagemaker_session = sagemaker.Session()\n",
    "predictor = Predictor(endpoint_name='deep-ar-100-bi', sagemaker_session=sagemaker_session)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 181,
   "metadata": {},
   "outputs": [],
   "source": [
    "data = {\n",
    "    \"instances\": [\n",
    "        {\n",
    "            \"start\": \"2024-08-31 00:00:00\",\n",
    "            \"target\": [657000.0, 693000.0],\n",
    "        }\n",
    "    ],\n",
    "    \"configuration\": {\n",
    "        \"num_samples\": 100,\n",
    "        \"output_types\": [\"mean\", \"quantiles\"],\n",
    "        \"quantiles\": [\"0.1\", \"0.5\", \"0.9\"]\n",
    "    }\n",
    "}\n",
    "\n",
    "payload = json.dumps(data)\n",
    "response = predictor.predict(payload, initial_args={\"ContentType\": \"application/json\"})"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 182,
   "metadata": {},
   "outputs": [],
   "source": [
    "result = json.loads(response.decode('utf-8'))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 198,
   "metadata": {},
   "outputs": [],
   "source": [
    "means = []\n",
    "means.append(nelson_bimonthly[-1:].apply(np.mean, axis=1).values[0])\n",
    "temp = [float(val) for val in result['predictions'][0]['mean']]\n",
    "means.append(temp[0])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 201,
   "metadata": {},
   "outputs": [],
   "source": [
    "index = [val for val in nelson_bimonthly.index]\n",
    "index.append(index[-1] + (pd.tseries.offsets.MonthEnd() * 2))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 204,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Text(0.5, 1.0, 'Forecast Horizon for Median House price in NZ')"
      ]
     },
     "execution_count": 204,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.plot(index[150:-1], nelson_bimonthly[150:].apply(np.mean, axis=1))\n",
    "plt.plot(index[-2:], means)\n",
    "plt.xlabel(\"Quarter\")\n",
    "plt.ylabel(\"Median House Price\")\n",
    "plt.title(\"Forecast Horizon for Median House price in NZ\")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Metrics"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 221,
   "metadata": {},
   "outputs": [],
   "source": [
    "def get_predictions(df):\n",
    "    predictions = []\n",
    "    for i in range(len(df)-1):\n",
    "        data = {\n",
    "            \"instances\": [\n",
    "                {\n",
    "                    \"start\": df.index[i].isoformat(),\n",
    "                    \"target\": [val for val in df.iloc[i]],\n",
    "                }\n",
    "            ],\n",
    "            \"configuration\": {\n",
    "                \"num_samples\": 100,\n",
    "                \"output_types\": [\"mean\", \"quantiles\"],\n",
    "                \"quantiles\": [\"0.1\", \"0.5\", \"0.9\"]\n",
    "            }\n",
    "        }\n",
    "\n",
    "        payload = json.dumps(data)\n",
    "        response = predictor.predict(payload, initial_args={\"ContentType\": \"application/json\"})\n",
    "        result = json.loads(response.decode('utf-8'))\n",
    "        pred = [float(val) for val in result['predictions'][0]['mean']]\n",
    "        \n",
    "        predictions.append(pred[0])\n",
    "        \n",
    "    return predictions"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 222,
   "metadata": {},
   "outputs": [],
   "source": [
    "predictions = get_predictions(nelson_bimonthly.copy())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 230,
   "metadata": {},
   "outputs": [],
   "source": [
    "predictions = pd.DataFrame(predictions, index=nelson_bimonthly.index[1:], columns=[\"predictions\"])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 234,
   "metadata": {},
   "outputs": [],
   "source": [
    "actual = nelson_bimonthly[1:].apply(np.mean, axis=1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 237,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "MSE: 546125232.9963598\n",
      "RMSE: 23369.322476194295\n",
      "MAE: 17212.46045673076\n",
      "MDAE: 12926.687500000116\n"
     ]
    }
   ],
   "source": [
    "mse = mean_squared_error(predictions.predictions, actual.values)\n",
    "rmse = root_mean_squared_error(predictions.predictions, actual.values)\n",
    "mae = mean_absolute_error(predictions.predictions, actual.values)\n",
    "mdae = median_absolute_error(predictions.predictions, actual.values)\n",
    "print(f'MSE: {mse}')\n",
    "print(f'RMSE: {rmse}')\n",
    "print(f'MAE: {mae}')\n",
    "print(f'MDAE: {mdae}')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 238,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[<matplotlib.lines.Line2D at 0x7f6ea877b250>]"
      ]
     },
     "execution_count": 238,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.plot(predictions.index, predictions.predictions)\n",
    "plt.plot(predictions.index, actual.values)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Although the metrics of this model are not the best, it's not hard to say that with some better hyperparameter tuning this algorithm could potentially fit the data very well. By using two monthly data the predicted value is reasonable, and the median absolute error (12926) is lower than some of my gradient boosting models, indicating great initial performance."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "---"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Conclusion"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "The two month data shows far better ability to predict the mean at an appropriate value. Furthermore, the prediction seems realistic and appears to predict a probably trend. The downside to this model is it can only extrapolate out two months, meaning the forecast horizon will only ever be two months into the future.\n",
    "\n",
    "It's unfortunate this model couldn't be used in deployment, or in our app for example, as it's possible train this algorithm further, or performing hyperparameter tuning could lead to a very good model."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "conda_python3",
   "language": "python",
   "name": "conda_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": "3.10.14"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 4
}