ver2.1_TCN.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "id": "96b5819f-e8ac-4c3c-9c28-93afb6ac738f",
   "metadata": {},
   "outputs": [],
   "source": [
    "import numpy as np\n",
    "import pandas as pd\n",
    "import torch\n",
    "import torch.nn as nn\n",
    "import torch.nn.functional as F\n",
    "import torch.optim as optim\n",
    "from torch.utils.data import TensorDataset, DataLoader\n",
    "from sklearn.preprocessing import MinMaxScaler, StandardScaler\n",
    "import matplotlib.pyplot as plt\n",
    "from tqdm import tqdm"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "id": "f788572f-61f5-4edd-a4b7-ba8c292c8589",
   "metadata": {
    "tags": []
   },
   "outputs": [],
   "source": [
    "##===\n",
    "# Select state we want to predict\n",
    "# New York , California , Florida , Texas\n",
    "STATE = 'Texas'\n",
    "LR = 5e-5\n",
    "WEIGHTDECAY = 0\n",
    "TRAIN = True\n",
    "seq_length = 7\n",
    "EPOCHS = 5000\n",
    "device = torch.device('cuda' if torch.cuda.is_available() else 'cpu')\n",
    "##===\n",
    "\n",
    "#### Model Config ===\n",
    "channel_config = np.array([[1,64,128],[128,64,64],[64,32,1]])"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "de64a227-a629-42e1-99b6-333734cca580",
   "metadata": {},
   "source": [
    "### Import Data from CSV file"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "id": "5e908914-f968-489e-a2a5-ef500dc4001f",
   "metadata": {},
   "outputs": [],
   "source": [
    "df = pd.read_csv('Cleaned_data.csv')"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "ad882fe1-7a80-4ca8-9a54-82dc294c48df",
   "metadata": {},
   "source": [
    "## Data frame Format"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "id": "344a4ff7-ad56-4738-93fa-6fadca83fe0f",
   "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>date</th>\n",
       "      <th>state</th>\n",
       "      <th>fips</th>\n",
       "      <th>cases</th>\n",
       "      <th>daily_cases</th>\n",
       "      <th>deaths</th>\n",
       "      <th>daily_deaths</th>\n",
       "      <th>total_vaccinations</th>\n",
       "      <th>total_distributed</th>\n",
       "      <th>daily_distributed</th>\n",
       "      <th>total_people_vaccinated</th>\n",
       "      <th>daily_people_vaccinated</th>\n",
       "      <th>total_people_fully_vaccinated</th>\n",
       "      <th>daily_people_fully_vaccinated</th>\n",
       "      <th>daily_vaccinations</th>\n",
       "      <th>mobility_composite</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>3/13/2021</td>\n",
       "      <td>Alabama</td>\n",
       "      <td>1</td>\n",
       "      <td>503305</td>\n",
       "      <td>594</td>\n",
       "      <td>10327</td>\n",
       "      <td>28</td>\n",
       "      <td>1288242</td>\n",
       "      <td>1918560.0</td>\n",
       "      <td>80160.0</td>\n",
       "      <td>841394.0</td>\n",
       "      <td>24976.0</td>\n",
       "      <td>492755.0</td>\n",
       "      <td>20430.0</td>\n",
       "      <td>26,935</td>\n",
       "      <td>-1.013041</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>3/12/2021</td>\n",
       "      <td>Alabama</td>\n",
       "      <td>1</td>\n",
       "      <td>502711</td>\n",
       "      <td>448</td>\n",
       "      <td>10299</td>\n",
       "      <td>25</td>\n",
       "      <td>1249205</td>\n",
       "      <td>1838400.0</td>\n",
       "      <td>22170.0</td>\n",
       "      <td>816418.0</td>\n",
       "      <td>25554.0</td>\n",
       "      <td>472325.0</td>\n",
       "      <td>19579.0</td>\n",
       "      <td>26,375</td>\n",
       "      <td>-1.013041</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>3/11/2021</td>\n",
       "      <td>Alabama</td>\n",
       "      <td>1</td>\n",
       "      <td>502263</td>\n",
       "      <td>865</td>\n",
       "      <td>10274</td>\n",
       "      <td>52</td>\n",
       "      <td>1209802</td>\n",
       "      <td>1816230.0</td>\n",
       "      <td>40570.0</td>\n",
       "      <td>790864.0</td>\n",
       "      <td>22362.0</td>\n",
       "      <td>452746.0</td>\n",
       "      <td>17728.0</td>\n",
       "      <td>25,376</td>\n",
       "      <td>-1.013041</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>3/10/2021</td>\n",
       "      <td>Alabama</td>\n",
       "      <td>1</td>\n",
       "      <td>501398</td>\n",
       "      <td>782</td>\n",
       "      <td>10222</td>\n",
       "      <td>36</td>\n",
       "      <td>1173820</td>\n",
       "      <td>1775660.0</td>\n",
       "      <td>86610.0</td>\n",
       "      <td>768502.0</td>\n",
       "      <td>21234.0</td>\n",
       "      <td>435018.0</td>\n",
       "      <td>14935.0</td>\n",
       "      <td>24,827</td>\n",
       "      <td>-1.013041</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>3/9/2021</td>\n",
       "      <td>Alabama</td>\n",
       "      <td>1</td>\n",
       "      <td>500616</td>\n",
       "      <td>524</td>\n",
       "      <td>10186</td>\n",
       "      <td>38</td>\n",
       "      <td>1140984</td>\n",
       "      <td>1689050.0</td>\n",
       "      <td>60700.0</td>\n",
       "      <td>747268.0</td>\n",
       "      <td>14301.0</td>\n",
       "      <td>420083.0</td>\n",
       "      <td>10146.0</td>\n",
       "      <td>24,118</td>\n",
       "      <td>-1.013041</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "        date    state  fips   cases  daily_cases  deaths  daily_deaths  \\\n",
       "0  3/13/2021  Alabama     1  503305          594   10327            28   \n",
       "1  3/12/2021  Alabama     1  502711          448   10299            25   \n",
       "2  3/11/2021  Alabama     1  502263          865   10274            52   \n",
       "3  3/10/2021  Alabama     1  501398          782   10222            36   \n",
       "4   3/9/2021  Alabama     1  500616          524   10186            38   \n",
       "\n",
       "  total_vaccinations  total_distributed  daily_distributed  \\\n",
       "0            1288242          1918560.0            80160.0   \n",
       "1            1249205          1838400.0            22170.0   \n",
       "2            1209802          1816230.0            40570.0   \n",
       "3            1173820          1775660.0            86610.0   \n",
       "4            1140984          1689050.0            60700.0   \n",
       "\n",
       "   total_people_vaccinated  daily_people_vaccinated  \\\n",
       "0                 841394.0                  24976.0   \n",
       "1                 816418.0                  25554.0   \n",
       "2                 790864.0                  22362.0   \n",
       "3                 768502.0                  21234.0   \n",
       "4                 747268.0                  14301.0   \n",
       "\n",
       "   total_people_fully_vaccinated  daily_people_fully_vaccinated  \\\n",
       "0                       492755.0                        20430.0   \n",
       "1                       472325.0                        19579.0   \n",
       "2                       452746.0                        17728.0   \n",
       "3                       435018.0                        14935.0   \n",
       "4                       420083.0                        10146.0   \n",
       "\n",
       "  daily_vaccinations  mobility_composite  \n",
       "0             26,935           -1.013041  \n",
       "1             26,375           -1.013041  \n",
       "2             25,376           -1.013041  \n",
       "3             24,827           -1.013041  \n",
       "4             24,118           -1.013041  "
      ]
     },
     "execution_count": 4,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "id": "21ae967e-fe62-4c78-a5e9-b3ce7158917c",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "<class 'pandas.core.frame.DataFrame'>\n",
      "RangeIndex: 20618 entries, 0 to 20617\n",
      "Data columns (total 16 columns):\n",
      " #   Column                         Non-Null Count  Dtype  \n",
      "---  ------                         --------------  -----  \n",
      " 0   date                           20618 non-null  object \n",
      " 1   state                          20618 non-null  object \n",
      " 2   fips                           20618 non-null  int64  \n",
      " 3   cases                          20618 non-null  int64  \n",
      " 4   daily_cases                    20618 non-null  int64  \n",
      " 5   deaths                         20618 non-null  int64  \n",
      " 6   daily_deaths                   20618 non-null  int64  \n",
      " 7   total_vaccinations             3801 non-null   object \n",
      " 8   total_distributed              3801 non-null   float64\n",
      " 9   daily_distributed              3801 non-null   float64\n",
      " 10  total_people_vaccinated        3801 non-null   float64\n",
      " 11  daily_people_vaccinated        3801 non-null   float64\n",
      " 12  total_people_fully_vaccinated  3801 non-null   float64\n",
      " 13  daily_people_fully_vaccinated  3801 non-null   float64\n",
      " 14  daily_vaccinations             3779 non-null   object \n",
      " 15  mobility_composite             19353 non-null  float64\n",
      "dtypes: float64(7), int64(5), object(4)\n",
      "memory usage: 2.5+ MB\n"
     ]
    }
   ],
   "source": [
    "df.info()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "id": "836c3722-619c-4ea8-a6f0-22214d8ee8b9",
   "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>date</th>\n",
       "      <th>state</th>\n",
       "      <th>fips</th>\n",
       "      <th>daily_cases</th>\n",
       "      <th>daily_deaths</th>\n",
       "      <th>daily_distributed</th>\n",
       "      <th>daily_people_vaccinated</th>\n",
       "      <th>daily_people_fully_vaccinated</th>\n",
       "      <th>daily_vaccinations</th>\n",
       "      <th>mobility_composite</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>3/13/2021</td>\n",
       "      <td>Alabama</td>\n",
       "      <td>1</td>\n",
       "      <td>594</td>\n",
       "      <td>28</td>\n",
       "      <td>80160.0</td>\n",
       "      <td>24976.0</td>\n",
       "      <td>20430.0</td>\n",
       "      <td>26,935</td>\n",
       "      <td>-1.013041</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>3/12/2021</td>\n",
       "      <td>Alabama</td>\n",
       "      <td>1</td>\n",
       "      <td>448</td>\n",
       "      <td>25</td>\n",
       "      <td>22170.0</td>\n",
       "      <td>25554.0</td>\n",
       "      <td>19579.0</td>\n",
       "      <td>26,375</td>\n",
       "      <td>-1.013041</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>3/11/2021</td>\n",
       "      <td>Alabama</td>\n",
       "      <td>1</td>\n",
       "      <td>865</td>\n",
       "      <td>52</td>\n",
       "      <td>40570.0</td>\n",
       "      <td>22362.0</td>\n",
       "      <td>17728.0</td>\n",
       "      <td>25,376</td>\n",
       "      <td>-1.013041</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>3/10/2021</td>\n",
       "      <td>Alabama</td>\n",
       "      <td>1</td>\n",
       "      <td>782</td>\n",
       "      <td>36</td>\n",
       "      <td>86610.0</td>\n",
       "      <td>21234.0</td>\n",
       "      <td>14935.0</td>\n",
       "      <td>24,827</td>\n",
       "      <td>-1.013041</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>3/9/2021</td>\n",
       "      <td>Alabama</td>\n",
       "      <td>1</td>\n",
       "      <td>524</td>\n",
       "      <td>38</td>\n",
       "      <td>60700.0</td>\n",
       "      <td>14301.0</td>\n",
       "      <td>10146.0</td>\n",
       "      <td>24,118</td>\n",
       "      <td>-1.013041</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "        date    state  fips  daily_cases  daily_deaths  daily_distributed  \\\n",
       "0  3/13/2021  Alabama     1          594            28            80160.0   \n",
       "1  3/12/2021  Alabama     1          448            25            22170.0   \n",
       "2  3/11/2021  Alabama     1          865            52            40570.0   \n",
       "3  3/10/2021  Alabama     1          782            36            86610.0   \n",
       "4   3/9/2021  Alabama     1          524            38            60700.0   \n",
       "\n",
       "   daily_people_vaccinated  daily_people_fully_vaccinated daily_vaccinations  \\\n",
       "0                  24976.0                        20430.0             26,935   \n",
       "1                  25554.0                        19579.0             26,375   \n",
       "2                  22362.0                        17728.0             25,376   \n",
       "3                  21234.0                        14935.0             24,827   \n",
       "4                  14301.0                        10146.0             24,118   \n",
       "\n",
       "   mobility_composite  \n",
       "0           -1.013041  \n",
       "1           -1.013041  \n",
       "2           -1.013041  \n",
       "3           -1.013041  \n",
       "4           -1.013041  "
      ]
     },
     "execution_count": 6,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df = df.drop(labels=['cases', 'deaths', 'total_vaccinations', 'total_distributed', 'total_people_vaccinated', 'total_people_fully_vaccinated'], axis=1)\n",
    "df.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "id": "cd39e90b-0daa-4780-a915-99e2c188b7f2",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array(['Alabama', 'Alaska', 'Arizona', 'Arkansas', 'California',\n",
       "       'Colorado', 'Connecticut', 'Delaware', 'District of Columbia',\n",
       "       'Florida', 'Georgia', 'Guam', 'Hawaii', 'Idaho', 'Illinois',\n",
       "       'Indiana', 'Iowa', 'Kansas', 'Kentucky', 'Louisiana', 'Maine',\n",
       "       'Maryland', 'Massachusetts', 'Michigan', 'Minnesota',\n",
       "       'Mississippi', 'Missouri', 'Montana', 'Nebraska', 'Nevada',\n",
       "       'New Hampshire', 'New Jersey', 'New Mexico', 'New York',\n",
       "       'North Carolina', 'North Dakota', 'Northern Mariana Islands',\n",
       "       'Ohio', 'Oklahoma', 'Oregon', 'Pennsylvania', 'Puerto Rico',\n",
       "       'Rhode Island', 'South Carolina', 'South Dakota', 'Tennessee',\n",
       "       'Texas', 'Utah', 'Vermont', 'Virgin Islands', 'Virginia',\n",
       "       'Washington', 'West Virginia', 'Wisconsin', 'Wyoming'],\n",
       "      dtype=object)"
      ]
     },
     "execution_count": 7,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df['state'].unique()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "id": "5d059c19-c413-4820-85a8-18632ef73b7f",
   "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>date</th>\n",
       "      <th>state</th>\n",
       "      <th>fips</th>\n",
       "      <th>daily_cases</th>\n",
       "      <th>daily_deaths</th>\n",
       "      <th>daily_distributed</th>\n",
       "      <th>daily_people_vaccinated</th>\n",
       "      <th>daily_people_fully_vaccinated</th>\n",
       "      <th>daily_vaccinations</th>\n",
       "      <th>mobility_composite</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>19484</th>\n",
       "      <td>2020-01-21</td>\n",
       "      <td>Washington</td>\n",
       "      <td>53</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>19483</th>\n",
       "      <td>2020-01-22</td>\n",
       "      <td>Washington</td>\n",
       "      <td>53</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>19482</th>\n",
       "      <td>2020-01-23</td>\n",
       "      <td>Washington</td>\n",
       "      <td>53</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5684</th>\n",
       "      <td>2020-01-24</td>\n",
       "      <td>Illinois</td>\n",
       "      <td>17</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>19481</th>\n",
       "      <td>2020-01-24</td>\n",
       "      <td>Washington</td>\n",
       "      <td>53</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</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",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>20250</th>\n",
       "      <td>2021-03-13</td>\n",
       "      <td>Wyoming</td>\n",
       "      <td>56</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>4086.0</td>\n",
       "      <td>2945.0</td>\n",
       "      <td>4,085</td>\n",
       "      <td>0.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4168</th>\n",
       "      <td>2021-03-13</td>\n",
       "      <td>Guam</td>\n",
       "      <td>66</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>226.0</td>\n",
       "      <td>1440.0</td>\n",
       "      <td>1,272</td>\n",
       "      <td>-5.361115</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>13475</th>\n",
       "      <td>2021-03-13</td>\n",
       "      <td>Northern Mariana Islands</td>\n",
       "      <td>69</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>230.0</td>\n",
       "      <td>163.0</td>\n",
       "      <td>246</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>15319</th>\n",
       "      <td>2021-03-13</td>\n",
       "      <td>Puerto Rico</td>\n",
       "      <td>72</td>\n",
       "      <td>235</td>\n",
       "      <td>0</td>\n",
       "      <td>19550.0</td>\n",
       "      <td>32455.0</td>\n",
       "      <td>20400.0</td>\n",
       "      <td>19,615</td>\n",
       "      <td>0.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>18330</th>\n",
       "      <td>2021-03-13</td>\n",
       "      <td>Virgin Islands</td>\n",
       "      <td>78</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>20618 rows × 10 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "            date                     state  fips  daily_cases  daily_deaths  \\\n",
       "19484 2020-01-21                Washington    53            1             0   \n",
       "19483 2020-01-22                Washington    53            0             0   \n",
       "19482 2020-01-23                Washington    53            0             0   \n",
       "5684  2020-01-24                  Illinois    17            1             0   \n",
       "19481 2020-01-24                Washington    53            0             0   \n",
       "...          ...                       ...   ...          ...           ...   \n",
       "20250 2021-03-13                   Wyoming    56            0             0   \n",
       "4168  2021-03-13                      Guam    66            1             0   \n",
       "13475 2021-03-13  Northern Mariana Islands    69            1             0   \n",
       "15319 2021-03-13               Puerto Rico    72          235             0   \n",
       "18330 2021-03-13            Virgin Islands    78            0             0   \n",
       "\n",
       "       daily_distributed  daily_people_vaccinated  \\\n",
       "19484                NaN                      NaN   \n",
       "19483                NaN                      NaN   \n",
       "19482                NaN                      NaN   \n",
       "5684                 NaN                      NaN   \n",
       "19481                NaN                      NaN   \n",
       "...                  ...                      ...   \n",
       "20250                0.0                   4086.0   \n",
       "4168                 0.0                    226.0   \n",
       "13475                0.0                    230.0   \n",
       "15319            19550.0                  32455.0   \n",
       "18330                NaN                      NaN   \n",
       "\n",
       "       daily_people_fully_vaccinated daily_vaccinations  mobility_composite  \n",
       "19484                            NaN                NaN                 NaN  \n",
       "19483                            NaN                NaN                 NaN  \n",
       "19482                            NaN                NaN                 NaN  \n",
       "5684                             NaN                NaN                 NaN  \n",
       "19481                            NaN                NaN                 NaN  \n",
       "...                              ...                ...                 ...  \n",
       "20250                         2945.0              4,085            0.000000  \n",
       "4168                          1440.0              1,272           -5.361115  \n",
       "13475                          163.0                246                 NaN  \n",
       "15319                        20400.0             19,615            0.000000  \n",
       "18330                            NaN                NaN                 NaN  \n",
       "\n",
       "[20618 rows x 10 columns]"
      ]
     },
     "execution_count": 8,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Set date to be datetime format and sort by date and fips\n",
    "df['date'] = pd.to_datetime(df.date)\n",
    "df = df.sort_values(by=['date', 'fips'], axis=0, ascending=True)\n",
    "df"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "id": "a6675f3e-4cdc-4c46-80ab-f222d50a4f86",
   "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>state</th>\n",
       "      <th>fips</th>\n",
       "      <th>daily_cases</th>\n",
       "      <th>daily_deaths</th>\n",
       "      <th>daily_distributed</th>\n",
       "      <th>daily_people_vaccinated</th>\n",
       "      <th>daily_people_fully_vaccinated</th>\n",
       "      <th>daily_vaccinations</th>\n",
       "      <th>mobility_composite</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>date</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>2020-02-12</th>\n",
       "      <td>Texas</td>\n",
       "      <td>48</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>-0.534251</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2020-02-13</th>\n",
       "      <td>Texas</td>\n",
       "      <td>48</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>-0.062539</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2020-02-14</th>\n",
       "      <td>Texas</td>\n",
       "      <td>48</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>0.373968</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2020-02-15</th>\n",
       "      <td>Texas</td>\n",
       "      <td>48</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>0.754066</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2020-02-16</th>\n",
       "      <td>Texas</td>\n",
       "      <td>48</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>1.069516</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",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2021-03-09</th>\n",
       "      <td>Texas</td>\n",
       "      <td>48</td>\n",
       "      <td>5409</td>\n",
       "      <td>184</td>\n",
       "      <td>444420.0</td>\n",
       "      <td>168326.0</td>\n",
       "      <td>68666.0</td>\n",
       "      <td>191,891</td>\n",
       "      <td>-7.686501</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2021-03-10</th>\n",
       "      <td>Texas</td>\n",
       "      <td>48</td>\n",
       "      <td>5350</td>\n",
       "      <td>202</td>\n",
       "      <td>815040.0</td>\n",
       "      <td>165705.0</td>\n",
       "      <td>77828.0</td>\n",
       "      <td>191,158</td>\n",
       "      <td>-7.686501</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2021-03-11</th>\n",
       "      <td>Texas</td>\n",
       "      <td>48</td>\n",
       "      <td>6185</td>\n",
       "      <td>212</td>\n",
       "      <td>150650.0</td>\n",
       "      <td>153209.0</td>\n",
       "      <td>60099.0</td>\n",
       "      <td>193,081</td>\n",
       "      <td>-7.686501</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2021-03-12</th>\n",
       "      <td>Texas</td>\n",
       "      <td>48</td>\n",
       "      <td>5230</td>\n",
       "      <td>223</td>\n",
       "      <td>225150.0</td>\n",
       "      <td>206533.0</td>\n",
       "      <td>92580.0</td>\n",
       "      <td>202,045</td>\n",
       "      <td>-7.738001</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2021-03-13</th>\n",
       "      <td>Texas</td>\n",
       "      <td>48</td>\n",
       "      <td>2876</td>\n",
       "      <td>141</td>\n",
       "      <td>81900.0</td>\n",
       "      <td>192733.0</td>\n",
       "      <td>85206.0</td>\n",
       "      <td>204,566</td>\n",
       "      <td>-7.738001</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>396 rows × 9 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "            state  fips  daily_cases  daily_deaths  daily_distributed  \\\n",
       "date                                                                    \n",
       "2020-02-12  Texas    48            1             0                NaN   \n",
       "2020-02-13  Texas    48            1             0                NaN   \n",
       "2020-02-14  Texas    48            0             0                NaN   \n",
       "2020-02-15  Texas    48            0             0                NaN   \n",
       "2020-02-16  Texas    48            0             0                NaN   \n",
       "...           ...   ...          ...           ...                ...   \n",
       "2021-03-09  Texas    48         5409           184           444420.0   \n",
       "2021-03-10  Texas    48         5350           202           815040.0   \n",
       "2021-03-11  Texas    48         6185           212           150650.0   \n",
       "2021-03-12  Texas    48         5230           223           225150.0   \n",
       "2021-03-13  Texas    48         2876           141            81900.0   \n",
       "\n",
       "            daily_people_vaccinated  daily_people_fully_vaccinated  \\\n",
       "date                                                                 \n",
       "2020-02-12                      NaN                            NaN   \n",
       "2020-02-13                      NaN                            NaN   \n",
       "2020-02-14                      NaN                            NaN   \n",
       "2020-02-15                      NaN                            NaN   \n",
       "2020-02-16                      NaN                            NaN   \n",
       "...                             ...                            ...   \n",
       "2021-03-09                 168326.0                        68666.0   \n",
       "2021-03-10                 165705.0                        77828.0   \n",
       "2021-03-11                 153209.0                        60099.0   \n",
       "2021-03-12                 206533.0                        92580.0   \n",
       "2021-03-13                 192733.0                        85206.0   \n",
       "\n",
       "           daily_vaccinations  mobility_composite  \n",
       "date                                               \n",
       "2020-02-12                NaN           -0.534251  \n",
       "2020-02-13                NaN           -0.062539  \n",
       "2020-02-14                NaN            0.373968  \n",
       "2020-02-15                NaN            0.754066  \n",
       "2020-02-16                NaN            1.069516  \n",
       "...                       ...                 ...  \n",
       "2021-03-09            191,891           -7.686501  \n",
       "2021-03-10            191,158           -7.686501  \n",
       "2021-03-11            193,081           -7.686501  \n",
       "2021-03-12            202,045           -7.738001  \n",
       "2021-03-13            204,566           -7.738001  \n",
       "\n",
       "[396 rows x 9 columns]"
      ]
     },
     "execution_count": 9,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df_clean = df[df.state == STATE]\n",
    "df_clean = df_clean.set_index('date')\n",
    "df_clean"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "id": "a8eccf57-7c24-4496-b2fa-2af17ebf8f28",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(378, 1)"
      ]
     },
     "execution_count": 10,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "rmlabels=['state', 'fips', 'daily_deaths', 'daily_distributed','daily_people_vaccinated', 'daily_people_fully_vaccinated',\n",
    "          'daily_vaccinations', 'mobility_composite']\n",
    "df_clean = df_clean.drop(labels=rmlabels, axis=1)\n",
    "df_clean = df_clean.iloc[df_clean.shape[0]-378:]\n",
    "df_clean.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "id": "c201a119-d9fa-4518-abef-46b721091fb8",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.axes._subplots.AxesSubplot at 0x7fdc2d6db9e8>"
      ]
     },
     "execution_count": 11,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7fdc2d6db048>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "df_clean.plot(title=f\"Daily cases :{STATE}\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "id": "23c4f150-3637-4570-adb3-6b5f40617af6",
   "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>date</th>\n",
       "      <th>state</th>\n",
       "      <th>fips</th>\n",
       "      <th>daily_cases</th>\n",
       "      <th>daily_deaths</th>\n",
       "      <th>daily_distributed</th>\n",
       "      <th>daily_people_vaccinated</th>\n",
       "      <th>daily_people_fully_vaccinated</th>\n",
       "      <th>daily_vaccinations</th>\n",
       "      <th>mobility_composite</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>19484</th>\n",
       "      <td>2020-01-21</td>\n",
       "      <td>Washington</td>\n",
       "      <td>53</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>19483</th>\n",
       "      <td>2020-01-22</td>\n",
       "      <td>Washington</td>\n",
       "      <td>53</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>19482</th>\n",
       "      <td>2020-01-23</td>\n",
       "      <td>Washington</td>\n",
       "      <td>53</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5684</th>\n",
       "      <td>2020-01-24</td>\n",
       "      <td>Illinois</td>\n",
       "      <td>17</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>19481</th>\n",
       "      <td>2020-01-24</td>\n",
       "      <td>Washington</td>\n",
       "      <td>53</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "            date       state  fips  daily_cases  daily_deaths  \\\n",
       "19484 2020-01-21  Washington    53            1             0   \n",
       "19483 2020-01-22  Washington    53            0             0   \n",
       "19482 2020-01-23  Washington    53            0             0   \n",
       "5684  2020-01-24    Illinois    17            1             0   \n",
       "19481 2020-01-24  Washington    53            0             0   \n",
       "\n",
       "       daily_distributed  daily_people_vaccinated  \\\n",
       "19484                NaN                      NaN   \n",
       "19483                NaN                      NaN   \n",
       "19482                NaN                      NaN   \n",
       "5684                 NaN                      NaN   \n",
       "19481                NaN                      NaN   \n",
       "\n",
       "       daily_people_fully_vaccinated daily_vaccinations  mobility_composite  \n",
       "19484                            NaN                NaN                 NaN  \n",
       "19483                            NaN                NaN                 NaN  \n",
       "19482                            NaN                NaN                 NaN  \n",
       "5684                             NaN                NaN                 NaN  \n",
       "19481                            NaN                NaN                 NaN  "
      ]
     },
     "execution_count": 12,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "id": "e8f7a48d-81a7-4b17-aedb-5923ffcbe32f",
   "metadata": {},
   "outputs": [],
   "source": [
    "df_state = df[df.state == STATE]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "id": "db9e99af-23dc-4e4e-a124-92d83a36d885",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7fdc2d463e10>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "import pandas as pd\n",
    "import seaborn as sn\n",
    "import matplotlib.pyplot as plt\n",
    "\n",
    "corrMatrix = df_state.corr()\n",
    "sn.heatmap(corrMatrix, annot=True)\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "6662892c-7d17-4a48-a17c-83731bd9759e",
   "metadata": {},
   "source": [
    "### select the high correlation to be our factors"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "a276e2d6-9e04-4b58-955b-a4631b50d829",
   "metadata": {},
   "source": [
    "- daily death\n",
    "- daily people vacinated \n",
    "- daily_distibuted  \n",
    "- daily case "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "id": "d6229abd-17d0-4e2b-b303-f4e2044f5079",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Index(['date', 'state', 'fips', 'daily_cases', 'daily_deaths',\n",
       "       'daily_distributed', 'daily_people_vaccinated',\n",
       "       'daily_people_fully_vaccinated', 'daily_vaccinations',\n",
       "       'mobility_composite'],\n",
       "      dtype='object')"
      ]
     },
     "execution_count": 15,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df_state.columns"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "id": "39b13fd3-a58b-476e-9bdd-65c1f96f8311",
   "metadata": {},
   "outputs": [],
   "source": [
    "df_state = df_state.set_index('date') "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "id": "598184b0-4082-404b-9b4d-3729f6297e39",
   "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>state</th>\n",
       "      <th>fips</th>\n",
       "      <th>daily_cases</th>\n",
       "      <th>daily_deaths</th>\n",
       "      <th>daily_distributed</th>\n",
       "      <th>daily_people_vaccinated</th>\n",
       "      <th>daily_people_fully_vaccinated</th>\n",
       "      <th>daily_vaccinations</th>\n",
       "      <th>mobility_composite</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>date</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>2020-02-12</th>\n",
       "      <td>Texas</td>\n",
       "      <td>48</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>-0.534251</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2020-02-13</th>\n",
       "      <td>Texas</td>\n",
       "      <td>48</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>-0.062539</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2020-02-14</th>\n",
       "      <td>Texas</td>\n",
       "      <td>48</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>0.373968</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2020-02-15</th>\n",
       "      <td>Texas</td>\n",
       "      <td>48</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>0.754066</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2020-02-16</th>\n",
       "      <td>Texas</td>\n",
       "      <td>48</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>1.069516</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "            state  fips  daily_cases  daily_deaths  daily_distributed  \\\n",
       "date                                                                    \n",
       "2020-02-12  Texas    48            1             0                NaN   \n",
       "2020-02-13  Texas    48            1             0                NaN   \n",
       "2020-02-14  Texas    48            0             0                NaN   \n",
       "2020-02-15  Texas    48            0             0                NaN   \n",
       "2020-02-16  Texas    48            0             0                NaN   \n",
       "\n",
       "            daily_people_vaccinated  daily_people_fully_vaccinated  \\\n",
       "date                                                                 \n",
       "2020-02-12                      NaN                            NaN   \n",
       "2020-02-13                      NaN                            NaN   \n",
       "2020-02-14                      NaN                            NaN   \n",
       "2020-02-15                      NaN                            NaN   \n",
       "2020-02-16                      NaN                            NaN   \n",
       "\n",
       "           daily_vaccinations  mobility_composite  \n",
       "date                                               \n",
       "2020-02-12                NaN           -0.534251  \n",
       "2020-02-13                NaN           -0.062539  \n",
       "2020-02-14                NaN            0.373968  \n",
       "2020-02-15                NaN            0.754066  \n",
       "2020-02-16                NaN            1.069516  "
      ]
     },
     "execution_count": 17,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df_state.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "id": "2bd8ef7b-99ba-4c1f-8c67-df4bf7ead2b7",
   "metadata": {},
   "outputs": [],
   "source": [
    "df_state = df_state.drop(['state', 'fips',\n",
    "       'daily_people_vaccinated',\n",
    "       'daily_people_fully_vaccinated',\n",
    "       'mobility_composite'], axis=1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "id": "8f2c8249-9da1-4cf9-b176-b70aa3610b4f",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "<class 'pandas.core.frame.DataFrame'>\n",
      "DatetimeIndex: 396 entries, 2020-02-12 to 2021-03-13\n",
      "Data columns (total 4 columns):\n",
      " #   Column              Non-Null Count  Dtype  \n",
      "---  ------              --------------  -----  \n",
      " 0   daily_cases         396 non-null    int64  \n",
      " 1   daily_deaths        396 non-null    int64  \n",
      " 2   daily_distributed   72 non-null     float64\n",
      " 3   daily_vaccinations  72 non-null     object \n",
      "dtypes: float64(1), int64(2), object(1)\n",
      "memory usage: 15.5+ KB\n"
     ]
    }
   ],
   "source": [
    "df_state.info()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "id": "d02afb67-eb2f-4da9-a7ab-75c97cc2cb93",
   "metadata": {},
   "outputs": [],
   "source": [
    "df_state_clean = df_state.dropna()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "id": "066ceb35-f5b7-47d2-ac4a-d36f27214d34",
   "metadata": {},
   "outputs": [],
   "source": [
    "df_state_clean = df_state_clean.replace(',','', regex=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "id": "3f9c5ecd-cbe0-45a4-9c6c-a9323ca98203",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(72, 4)"
      ]
     },
     "execution_count": 22,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df_state_clean.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "id": "ae9bb82b-3168-4ac9-be30-f19c39c8ace6",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "<class 'pandas.core.frame.DataFrame'>\n",
      "DatetimeIndex: 72 entries, 2021-01-01 to 2021-03-13\n",
      "Data columns (total 4 columns):\n",
      " #   Column              Non-Null Count  Dtype  \n",
      "---  ------              --------------  -----  \n",
      " 0   daily_cases         72 non-null     int64  \n",
      " 1   daily_deaths        72 non-null     int64  \n",
      " 2   daily_distributed   72 non-null     float64\n",
      " 3   daily_vaccinations  72 non-null     object \n",
      "dtypes: float64(1), int64(2), object(1)\n",
      "memory usage: 2.8+ KB\n"
     ]
    }
   ],
   "source": [
    "df_state_clean.info()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "id": "c2a67ca0-bb2c-45ba-ad06-818b775b284f",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(72, 4)"
      ]
     },
     "execution_count": 24,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df_state_clean.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "id": "e957bba3-ef61-40d2-bb90-0db349eb26e0",
   "metadata": {},
   "outputs": [],
   "source": [
    "def create_sequences(data, seq_length):\n",
    "    xs = []\n",
    "    ys = []\n",
    "    \n",
    "    for i in range(len(data)-seq_length-1):\n",
    "        x = data[i:(i+seq_length)]\n",
    "        #print(i)\n",
    "        y = data[i+seq_length]\n",
    "        xs.append(x)\n",
    "        ys.append(y)\n",
    "\n",
    "    return np.array(xs), np.array(ys)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "17a936f1-f68a-411a-b61c-c4067a1207bd",
   "metadata": {},
   "source": [
    "# Define Model"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "id": "ff4391ed-a40f-4fcd-9e6a-ead688102775",
   "metadata": {},
   "outputs": [],
   "source": [
    "class ResidualBlock(nn.Module):\n",
    "    def __init__(self,kernel_size=3,dilation=1,in_dim=16,hidden_dim=64,out_dim=128):\n",
    "        super(ResidualBlock, self).__init__()\n",
    "        self.pad = (kernel_size - 1) * dilation\n",
    "        \n",
    "        #self.casual1 = nn.Conv1d(in_channels=16,out_channels=32,kernel_size=kernel_size,stride=1,padding=self.pad,dilation=dilation)\n",
    "        #self.casual2 = nn.Conv1d(in_channels=32,out_channels=64,kernel_size=kernel_size,stride=1,padding=self.pad,dilation=dilation)\n",
    "        \n",
    "        self.layer_1 = nn.Sequential(nn.utils.weight_norm(nn.Conv1d(in_channels=in_dim,out_channels=hidden_dim,\n",
    "                                     kernel_size=kernel_size,stride=1,padding=self.pad,dilation=dilation)),\n",
    "                                     nn.LeakyReLU(0.1),nn.Dropout(p=0.2))\n",
    "         \n",
    "        self.layer_2 = nn.Sequential(nn.utils.weight_norm(nn.Conv1d(in_channels=hidden_dim,out_channels=out_dim,\n",
    "                                    kernel_size=kernel_size,stride=1,padding=self.pad,dilation=dilation)),\n",
    "                                    nn.LeakyReLU(0.1),nn.Dropout(p=0.2))\n",
    "        \n",
    "        self.conv1_1 = nn.Conv1d(in_channels=in_dim,out_channels=out_dim,\n",
    "                                                        kernel_size=1,stride=1,padding=0,dilation=1)\n",
    "        \n",
    "    def forward(self,x):\n",
    "        \n",
    "        x_iden = x\n",
    "        x = self.layer_1(x)\n",
    "        x = x[:,:,:-self.pad]\n",
    "        #print(x.shape)\n",
    "        x = self.layer_2(x)\n",
    "        x = x[:,:,:-self.pad]\n",
    "        #print(x.shape)\n",
    "        x_iden = self.conv1_1(x_iden)\n",
    "        #print(x_iden.shape)\n",
    "        x = x + x_iden\n",
    "        \n",
    "        return x         "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 50,
   "id": "0c50375b-7e32-4447-985b-3cb7c4a9a88f",
   "metadata": {},
   "outputs": [],
   "source": [
    "class TemporalConvNet(nn.Module):\n",
    "    def __init__(self,num_residual_layers,channel_config,num_classes):\n",
    "        super(TemporalConvNet, self).__init__()\n",
    "        self.num_residual_layers = num_residual_layers \n",
    "        self.residual_layers = []\n",
    "        #print(self.num_residual_layers)\n",
    "        \n",
    "        for i in range(self.num_residual_layers):\n",
    "            d = pow(2, i)\n",
    "            #print(d)\n",
    "            self.residual_layers.append(ResidualBlock(kernel_size=3,dilation=d,in_dim=channel_config[i][0],hidden_dim=channel_config[i][1],out_dim=channel_config[i][2]).to(device))        \n",
    "        #print(self.residual_layers)\n",
    "        \n",
    "        #1200\n",
    "        \n",
    "#         self.fc = nn.Sequential(nn.Linear(7 , 7),nn.LeakyReLU(0.1),nn.BatchNorm1d(num_features=7),nn.Dropout(0.5),\n",
    "#                                 nn.Linear(7,4),nn.LeakyReLU(0.1),nn.BatchNorm1d(num_features=4),\n",
    "#                                 nn.Dropout(0.5),nn.Linear(4,2),nn.BatchNorm1d(num_features=2))\n",
    "        \n",
    "        self.fc = nn.Sequential(nn.Linear(7 , 7),nn.LeakyReLU(0.1),nn.Dropout(0.5),\n",
    "                                nn.Linear(7,4),nn.LeakyReLU(0.1),\n",
    "                                nn.Dropout(0.5),nn.Linear(4,2))\n",
    "        \n",
    "        \n",
    "        self.out = nn.Linear(2, 1)\n",
    "        #self.softmax = nn.LogSoftmax(dim=1)\n",
    "        \n",
    "\n",
    "        #self.fc = nn.Sequential(nn.Linear(75 ,100),nn.ReLU(),nn.Linear(100,10))\n",
    "        \n",
    "    def forward(self,x):\n",
    "        for layer in self.residual_layers:\n",
    "            x = layer(x)\n",
    "        #print(\"layer :\",x.shape)\n",
    "        x = torch.flatten(x,start_dim=1)\n",
    "        #print(\"flatten:\",x.shape)        \n",
    "        x = self.fc(x).to(device)\n",
    "        \n",
    "        x = self.out(x).to(device)\n",
    "        #print(x.shape)\n",
    "    \n",
    "        return x"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 51,
   "id": "4828ea44-9f63-4640-a5d9-36de8a59dadb",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Configured device:  cuda\n"
     ]
    }
   ],
   "source": [
    "device = torch.device('cuda' if torch.cuda.is_available() else 'cpu')\n",
    "print(\"Configured device: \", device)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "41e6cf0c-dc31-46f8-9fb7-58a5f722b13d",
   "metadata": {},
   "source": [
    "## DATA Shape\n",
    "- (batch_size,data point,channel)\n",
    "\n",
    "### For CNN shape\n",
    "- Shape:\n",
    "        Input:(batch_size,channel, data point)\n",
    "        "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 52,
   "id": "6c184b07-5c84-4dfc-83c8-3ee084b21dc0",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.axes._subplots.AxesSubplot at 0x7fdc2d76d400>"
      ]
     },
     "execution_count": 52,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7fdc2cc31588>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "df_clean.plot(title=f\"Daily cases :{STATE}\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 53,
   "id": "ebbadcba-a484-49b3-93cd-4ecab2707ab3",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "378"
      ]
     },
     "execution_count": 53,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df_clean.shape[0]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 104,
   "id": "b1d6b808-6d67-4061-974b-d11b5c846246",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "266\n",
      "75\n",
      "37\n"
     ]
    }
   ],
   "source": [
    "train_size = int(len(df_clean)*0.7)+2\n",
    "valid_size = int(len(df_clean)*0.2)\n",
    "test_size  = int(len(df_clean)*0.1) \n",
    "\n",
    "print(train_size)\n",
    "print(valid_size)\n",
    "print(test_size)\n",
    "\n",
    "#print(train_size+valid_size+test_size)\n",
    "\n",
    "assert train_size +  valid_size + test_size == df_clean.shape[0]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 105,
   "id": "3c12e985-d837-4f1d-994f-09c7549ee47a",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "((266, 1), (37, 1), (75, 1))"
      ]
     },
     "execution_count": 105,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "train_data = df_clean.iloc[:train_size]\n",
    "test_data = df_clean.iloc[train_size:train_size+test_size]\n",
    "valid_data = df_clean.iloc[train_size+test_size:]\n",
    "\n",
    "\n",
    "\n",
    "# train_data = np.expand_dims(train_data, axis=1)\n",
    "# test_data = np.expand_dims(test_data, axis=1)\n",
    "\n",
    "train_data.shape, test_data.shape , valid_data.shape "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 106,
   "id": "9f83ef7b-d478-42f6-ae8e-f5c3a3a8239f",
   "metadata": {},
   "outputs": [],
   "source": [
    "# scaler = MinMaxScaler()\n",
    "scaler = StandardScaler()\n",
    "\n",
    "scaler = scaler.fit(train_data)\n",
    "\n",
    "train_data = scaler.transform(train_data)\n",
    "\n",
    "test_data = scaler.transform(test_data)\n",
    "\n",
    "valid_data = scaler.transform(valid_data)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 107,
   "id": "15532477-1a0e-44c2-a270-128a09cefaab",
   "metadata": {},
   "outputs": [],
   "source": [
    "train_data = np.asarray(train_data)\n",
    "test_data = np.asarray(test_data)\n",
    "valid_data = np.asarray(valid_data)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 108,
   "id": "580acfdd-8af4-470d-8b41-b4a737e38cbc",
   "metadata": {},
   "outputs": [],
   "source": [
    "X_train, y_train = create_sequences(train_data, seq_length)\n",
    "X_test, y_test = create_sequences(test_data, seq_length)\n",
    "X_valid ,y_valid = create_sequences(valid_data, seq_length)\n",
    "\n",
    "X_train = torch.from_numpy(X_train).float()\n",
    "y_train = torch.from_numpy(y_train).float()\n",
    "\n",
    "X_test = torch.from_numpy(X_test).float()\n",
    "y_test = torch.from_numpy(y_test).float()\n",
    "\n",
    "\n",
    "X_valid = torch.from_numpy(X_valid).float()\n",
    "y_valid = torch.from_numpy(y_valid).float()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 109,
   "id": "6a9dcc9e-4638-48f6-859b-09d860477355",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(torch.Size([258, 7, 1]),\n",
       " torch.Size([258, 1]),\n",
       " torch.Size([29, 7, 1]),\n",
       " torch.Size([29, 1]))"
      ]
     },
     "execution_count": 109,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "X_train.shape , y_train.shape , X_test.shape, y_test.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 110,
   "id": "cdc119a6-14a0-4285-ab6e-3bdf3b9b3f9a",
   "metadata": {},
   "outputs": [],
   "source": [
    "X_train = torch.reshape(X_train,(X_train.shape[0],1,-1))\n",
    "X_test =  torch.reshape(X_test,(X_test.shape[0],1,-1))\n",
    "X_valid =  torch.reshape(X_valid,(X_valid.shape[0],1,-1))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 111,
   "id": "966746ba-33e5-4424-8346-37dd3c7bea01",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(torch.Size([258, 1, 7]),\n",
       " torch.Size([258, 1]),\n",
       " torch.Size([29, 1, 7]),\n",
       " torch.Size([29, 1]))"
      ]
     },
     "execution_count": 111,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "X_train.shape , y_train.shape , X_test.shape, y_test.shape"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "de6c5fde-56f8-48a3-ad70-2e28dbb96c4d",
   "metadata": {},
   "source": [
    "### Data loader"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 112,
   "id": "365d5788-cc34-4817-a2e7-bd0811710ee6",
   "metadata": {},
   "outputs": [],
   "source": [
    "from torch.utils.data import TensorDataset\n",
    "\n",
    "# Define dataset\n",
    "ds_train = TensorDataset(X_train, y_train)\n",
    "ds_test = TensorDataset(X_test, y_test)\n",
    "ds_valid = TensorDataset(X_valid, y_valid)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 113,
   "id": "30d22c12-cb30-4a04-9280-85fa7018d82a",
   "metadata": {},
   "outputs": [],
   "source": [
    "BATCH_SIZE = 7  #keeping it binary so it fits GPU\n",
    "# #Train set loader\n",
    "train_iterator = torch.utils.data.DataLoader(dataset=ds_train, \n",
    "                                           batch_size=BATCH_SIZE, \n",
    "                                            shuffle=True)\n",
    "#Test set loader\n",
    "test_iterator = torch.utils.data.DataLoader(dataset=ds_test, \n",
    "                                          batch_size=BATCH_SIZE, \n",
    "                                          shuffle=True)\n",
    "\n",
    "valid_iterator = torch.utils.data.DataLoader(dataset=ds_valid, \n",
    "                                          batch_size=BATCH_SIZE, \n",
    "                                          shuffle=True)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "4b5b241f-00f7-4bf2-9c21-8852104cfc83",
   "metadata": {},
   "source": [
    "### Create model"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 64,
   "id": "49e276f0-ef26-49fe-982e-9846df90261c",
   "metadata": {},
   "outputs": [],
   "source": [
    "model_Tcn = TemporalConvNet(num_residual_layers=3,channel_config=channel_config,num_classes=1)\n",
    "model_Tcn = model_Tcn.float()\n",
    "models = [model_Tcn]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 65,
   "id": "e9b02cf6-e8f4-4885-844c-fbe367ee456c",
   "metadata": {},
   "outputs": [],
   "source": [
    "for model in models:\n",
    "    model = model.to(device)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "cf3c0d2e-9240-4b30-8695-b8c86e02506c",
   "metadata": {},
   "source": [
    "### Create Optimizer"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 66,
   "id": "ed97a139-94a0-48b7-b415-ccf51a58b346",
   "metadata": {},
   "outputs": [],
   "source": [
    "import torch.optim as optim\n",
    "#0.01\n",
    "learning_rate = 1e-7\n",
    "\n",
    "optimizer = optim.Adam(model.parameters())\n",
    "\n",
    "criterion = torch.nn.MSELoss() #nn.NLLLoss()\n",
    "optimizer = torch.optim.Adam(model.parameters(), lr=learning_rate)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "5fe76e97-cd26-4824-8bb1-1b7ca11631fd",
   "metadata": {},
   "source": [
    "### Training Function"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 70,
   "id": "a2b7a551-3e07-4e10-9547-c0d499e5a9a6",
   "metadata": {},
   "outputs": [],
   "source": [
    "def train(model, iterator, optimizer, criterion):\n",
    "    total = 0\n",
    "    correct = 0\n",
    "    epoch_loss = 0\n",
    "    epoch_acc = 0\n",
    "   # loss_list = []\n",
    "    model.train()\n",
    "    \n",
    "    for batch, num_cases in iterator:\n",
    "        \n",
    "        #Move tensors to the configured device\n",
    "        batch = batch.to(device)\n",
    "        num_cases = num_cases.to(device) \n",
    "#         num_cases = num_cases[:,-1]\n",
    "#         num_cases =  torch.reshape(num_cases, (num_cases.shape[0], 1))\n",
    "        #Forward pass\n",
    "        #print(\"batch train:\",batch.shape)\n",
    "        outputs = model(batch.float())\n",
    "#         print(\"outputs:\",outputs.shape)\n",
    "#         print(\"num_cases :\",num_cases.shape)\n",
    "#         print(\"outputs :\",outputs)\n",
    "#         print(\"num_cases :\",num_cases)\n",
    "        loss = criterion(outputs, num_cases.float())\n",
    "        \n",
    "        # Backward and optimize\n",
    "        optimizer.zero_grad()\n",
    "        loss.backward()\n",
    "        optimizer.step()\n",
    "        #loss_list.append(loss.data)  \n",
    "         \n",
    "#         #check accuracy\n",
    "#         predictions = model(batch.float())\n",
    "#         _, predicted = torch.max(predictions.data, 1)  #returns max value, indices\n",
    "#          #keep track of total\n",
    "#         correct += (predicted == labels).sum().item()  #.item() give the raw number\n",
    "#         acc = 100 * (correct / total)\n",
    "                \n",
    "        epoch_loss += loss.item()\n",
    "#         epoch_acc = acc\n",
    "      \n",
    "    return epoch_loss / len(iterator)  #, epoch_acc, predicted"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 71,
   "id": "d816be15-ba87-40e5-8182-8983d08fb978",
   "metadata": {},
   "outputs": [],
   "source": [
    "def test(model,iterator, optimizer, criterion):\n",
    "    total = 0\n",
    "    correct = 0\n",
    "    epoch_loss = 0\n",
    "    epoch_acc = 0\n",
    "   # loss_list = []\n",
    "    for batch, num_cases in iterator:\n",
    "        batch = batch.to(device)\n",
    "        num_cases = num_cases.to(device)\n",
    "        #print(\"batch test:\",batch.shape)\n",
    "        with torch.no_grad():\n",
    "            y_test_pred = model(batch.float())\n",
    "            loss = criterion(y_test_pred, num_cases.float())\n",
    "            epoch_loss += loss\n",
    "            \n",
    "    return epoch_loss / len(iterator)  "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 77,
   "id": "e4afaa15-4100-4c52-80fb-c7ca85fbe37c",
   "metadata": {},
   "outputs": [],
   "source": [
    "from IPython.display import clear_output"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 78,
   "id": "66e0a334-0206-4e81-ac46-7a8b1c92fcfb",
   "metadata": {},
   "outputs": [],
   "source": [
    "N_EPOCHS = 5000\n",
    "name = f\"Epoch ={N_EPOCHS} LR = {learning_rate} Adam\""
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 80,
   "id": "9509ba65-c59a-4318-886e-f7e0e565770a",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 4951 train loss: 1.099078142159694 test loss: 11.809953689575195\n",
      "Epoch 4952 train loss: 1.1013113115284894 test loss: 45.68912124633789\n",
      "Epoch 4953 train loss: 1.101764650763692 test loss: 14.650690078735352\n",
      "Epoch 4954 train loss: 1.101865249711114 test loss: 12.206536293029785\n",
      "Epoch 4955 train loss: 1.1011831047567162 test loss: 12.267387390136719\n",
      "Epoch 4956 train loss: 1.0989806748725273 test loss: 11.415916442871094\n",
      "Epoch 4957 train loss: 1.09795688132982 test loss: 11.540812492370605\n",
      "Epoch 4958 train loss: 1.1025716034141746 test loss: 12.916511535644531\n",
      "Epoch 4959 train loss: 1.1005442754642383 test loss: 11.419063568115234\n",
      "Epoch 4960 train loss: 1.1021283526678343 test loss: 45.55609893798828\n",
      "Epoch 4961 train loss: 1.0990194476939537 test loss: 14.179316520690918\n",
      "Epoch 4962 train loss: 1.0979095794059135 test loss: 13.208670616149902\n",
      "Epoch 4963 train loss: 1.1059614887108673 test loss: 11.435333251953125\n",
      "Epoch 4964 train loss: 1.1023642187182967 test loss: 45.57771682739258\n",
      "Epoch 4965 train loss: 1.1005967339953862 test loss: 12.494772911071777\n",
      "Epoch 4966 train loss: 1.0991315366448582 test loss: 14.702423095703125\n",
      "Epoch 4967 train loss: 1.0999997234022296 test loss: 11.758816719055176\n",
      "Epoch 4968 train loss: 1.1022072345823855 test loss: 14.620535850524902\n",
      "Epoch 4969 train loss: 1.0979938321822398 test loss: 11.787654876708984\n",
      "Epoch 4970 train loss: 1.102397751163792 test loss: 12.243025779724121\n",
      "Epoch 4971 train loss: 1.104532791150583 test loss: 12.048747062683105\n",
      "Epoch 4972 train loss: 1.1050012667436857 test loss: 11.966570854187012\n",
      "Epoch 4973 train loss: 1.1019326963940181 test loss: 11.73535442352295\n",
      "Epoch 4974 train loss: 1.1009047442191355 test loss: 12.126042366027832\n",
      "Epoch 4975 train loss: 1.1033527637655671 test loss: 11.77291202545166\n",
      "Epoch 4976 train loss: 1.1011464257498045 test loss: 11.769906997680664\n",
      "Epoch 4977 train loss: 1.098811388015747 test loss: 13.723215103149414\n",
      "Epoch 4978 train loss: 1.1038561592230927 test loss: 11.766630172729492\n",
      "Epoch 4979 train loss: 1.1003491387173936 test loss: 11.3994722366333\n",
      "Epoch 4980 train loss: 1.0988069192783252 test loss: 13.218174934387207\n",
      "Epoch 4981 train loss: 1.101669711035651 test loss: 14.14258098602295\n",
      "Epoch 4982 train loss: 1.0994469056258331 test loss: 12.942988395690918\n",
      "Epoch 4983 train loss: 1.1003843690897968 test loss: 11.436652183532715\n",
      "Epoch 4984 train loss: 1.102560874578115 test loss: 11.47080135345459\n",
      "Epoch 4985 train loss: 1.1035891569949485 test loss: 11.384708404541016\n",
      "Epoch 4986 train loss: 1.0980097168200724 test loss: 13.197656631469727\n",
      "Epoch 4987 train loss: 1.1003023361837543 test loss: 45.67833709716797\n",
      "Epoch 4988 train loss: 1.1011172718292959 test loss: 45.557804107666016\n",
      "Epoch 4989 train loss: 1.1016885146901414 test loss: 14.929450035095215\n",
      "Epoch 4990 train loss: 1.1194466698814083 test loss: 12.239908218383789\n",
      "Epoch 4991 train loss: 1.102768651536993 test loss: 12.995340347290039\n",
      "Epoch 4992 train loss: 1.1038914294661701 test loss: 12.229351043701172\n",
      "Epoch 4993 train loss: 1.099631583368456 test loss: 11.452702522277832\n",
      "Epoch 4994 train loss: 1.105074194637505 test loss: 13.754765510559082\n",
      "Epoch 4995 train loss: 1.1054231468084696 test loss: 11.536998748779297\n",
      "Epoch 4996 train loss: 1.1042875679763589 test loss: 11.45313835144043\n",
      "Epoch 4997 train loss: 1.1021760677969135 test loss: 12.243304252624512\n",
      "Epoch 4998 train loss: 1.0997079143653046 test loss: 12.073328018188477\n",
      "Epoch 4999 train loss: 1.1040156467540845 test loss: 11.789392471313477\n"
     ]
    }
   ],
   "source": [
    "\n",
    "best_valid_loss = float('inf')\n",
    "loss_train = []\n",
    "loss_test= []\n",
    "for i, model in enumerate(models):\n",
    "    print(f\"Training {type(model).__name__}\")\n",
    "    \n",
    "    #start_time = time.time()\n",
    "\n",
    "    for epoch in range(N_EPOCHS):\n",
    "        #start_time = time.time()\n",
    "        loss  = train(model, train_iterator, optimizer, criterion) \n",
    "        test_loss = test(model, test_iterator, optimizer, criterion)\n",
    "        print(f'Epoch {epoch} train loss: {loss} test loss: {test_loss}') \n",
    "        loss_train.append(loss)\n",
    "        loss_test.append(test_loss)\n",
    "        \n",
    "        if epoch % 50 == 0:\n",
    "            clear_output(wait=True)\n",
    "            \n",
    "        if test_loss  < best_valid_loss:\n",
    "                best_valid_loss = test_loss\n",
    "                torch.save(model.state_dict(), f'Models/{name}.pt')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "ee71c1b6-300c-46b8-96f2-911ffdbbc47a",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": 81,
   "id": "72a1cc4c-8aa1-4d0e-8519-bee36c293b40",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.legend.Legend at 0x7fdc0c652a20>"
      ]
     },
     "execution_count": 81,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7fdc0c7460b8>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.plot(loss_train)\n",
    "plt.plot(loss_test)\n",
    "plt.legend(['loss_train','loss_test'])"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "4b36b73e-bad8-43fd-b70f-66af1b62b0c5",
   "metadata": {},
   "source": [
    "## "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 142,
   "id": "118f2c99-361d-4084-ab4c-f82afc34c2eb",
   "metadata": {},
   "outputs": [],
   "source": [
    "yhat = []\n",
    "ytrue = []\n",
    "\n",
    "test_loss = 0\n",
    "for i, model in enumerate(models):\n",
    "    with torch.no_grad(): \n",
    "        for batch_test, num_cases_test in valid_iterator:\n",
    "            batch_test = batch_test.to(device)\n",
    "            num_cases = num_cases_test.to(device)\n",
    "            outputs = model(batch_test)\n",
    "            \n",
    "#             outputs = outputs    \n",
    "#             num_cases = num_cases \n",
    "#             outputs = scaler.inverse_transform(outputs)\n",
    "#             num_cases = scaler.inverse_transform(num_cases)\n",
    "                            \n",
    "            prediction = torch.flatten(outputs)\n",
    "            real_valid = torch.flatten(num_cases_test)\n",
    "                            \n",
    "            prediction = scaler.inverse_transform(prediction.cpu())\n",
    "            real_valid = scaler.inverse_transform(real_valid.cpu())\n",
    "            \n",
    "            yhat.append(prediction.tolist())\n",
    "            ytrue.append(real_valid.tolist())\n",
    "#             print(prediction)\n",
    "#             print(real_valid)\n",
    "            "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "25078c02-a8be-4f1e-a706-2ec7cd3c8a27",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": 143,
   "id": "38ebdf50-f85d-4aac-844d-7c073681c1bf",
   "metadata": {},
   "outputs": [],
   "source": [
    "from itertools import chain"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 147,
   "id": "7b7c493c-5a54-478c-b51c-3e23a3af062d",
   "metadata": {},
   "outputs": [
    {
     "ename": "TypeError",
     "evalue": "'float' object is not iterable",
     "output_type": "error",
     "traceback": [
      "\u001b[0;31m---------------------------------------------------------------------------\u001b[0m",
      "\u001b[0;31mTypeError\u001b[0m                                 Traceback (most recent call last)",
      "\u001b[0;32m<ipython-input-147-0f14d072b547>\u001b[0m in \u001b[0;36m<module>\u001b[0;34m\u001b[0m\n\u001b[0;32m----> 1\u001b[0;31m \u001b[0myhat\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mlist\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mchain\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mfrom_iterable\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0myhat\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m      2\u001b[0m \u001b[0mytrue\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mlist\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mchain\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mfrom_iterable\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mytrue\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;31mTypeError\u001b[0m: 'float' object is not iterable"
     ]
    }
   ],
   "source": [
    "yhat = list(chain.from_iterable(yhat))\n",
    "ytrue = list(chain.from_iterable(ytrue))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 146,
   "id": "aaa1f582-7157-4bc8-a234-03cc3fcfeeaa",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": 149,
   "id": "952f06cd-9c36-4b11-b12b-9c3168a79746",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.legend.Legend at 0x7fdcfce10d68>"
      ]
     },
     "execution_count": 149,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7fdcfcbd8b38>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.plot(yhat)\n",
    "plt.plot(ytrue)\n",
    "plt.legend(['yhat','ytrue'])"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "a4a027dd-ba91-4b82-b01e-048a6c0c0537",
   "metadata": {},
   "source": [
    "# Todo\n",
    "1) check real value of validation before and after standard scaler\n",
    "2) "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "b02cd107-8bb1-4314-a073-77ae8879bc28",
   "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": "3.6.9"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}