Tensorflow - Complete Tensorflow 2 and Keras Deep Learning Bootcamp - JP/05-RNNs/04-Multivariate-Time-Series-RNN.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Multivariate Time Series with RNN\n",
    "\n",
    "\n",
    "## NOTE: In general, it is very difficult to get good performance from an LSTM based RNN model for multivariate series. We highly recommend using SARIMAX based methods if you have a real-world situation requiring multi-variate time series data.\n",
    "\n",
    "---\n",
    "Appliance Energy Usage: A Multivariate Time Series Forecasting Example\n",
    "\n",
    "Experimental data used to create regression models of appliances energy use in a low energy building.\n",
    "Data Set Information:\n",
    "\n",
    "The data set is at 10 min for about 4.5 months. The house temperature and humidity conditions were monitored with a ZigBee wireless sensor network. Each wireless node transmitted the temperature and humidity conditions around 3.3 min. Then, the wireless data was averaged for 10 minutes periods. The energy data was logged every 10 minutes with m-bus energy meters. Weather from the nearest airport weather station (Chievres Airport, Belgium) was downloaded from a public data set from Reliable Prognosis (rp5.ru), and merged together with the experimental data sets using the date and time column. Two random variables have been included in the data set for testing the regression models and to filter out non predictive attributes (parameters).\n",
    "Original source of the dataset:\n",
    "\n",
    "http://archive.ics.uci.edu/ml/datasets/Appliances+energy+prediction\n",
    "---"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "import numpy as np\n",
    "import pandas as pd\n",
    "import matplotlib.pyplot as plt\n",
    "import seaborn as sns"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [],
   "source": [
    "df = pd.read_csv('../Data/energydata_complete.csv', index_col='date', infer_datetime_format=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "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>Appliances</th>\n",
       "      <th>lights</th>\n",
       "      <th>T1</th>\n",
       "      <th>RH_1</th>\n",
       "      <th>T2</th>\n",
       "      <th>RH_2</th>\n",
       "      <th>T3</th>\n",
       "      <th>RH_3</th>\n",
       "      <th>T4</th>\n",
       "      <th>RH_4</th>\n",
       "      <th>...</th>\n",
       "      <th>T9</th>\n",
       "      <th>RH_9</th>\n",
       "      <th>T_out</th>\n",
       "      <th>Press_mm_hg</th>\n",
       "      <th>RH_out</th>\n",
       "      <th>Windspeed</th>\n",
       "      <th>Visibility</th>\n",
       "      <th>Tdewpoint</th>\n",
       "      <th>rv1</th>\n",
       "      <th>rv2</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",
       "      <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",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>2016-01-11 17:00:00</th>\n",
       "      <td>60</td>\n",
       "      <td>30</td>\n",
       "      <td>19.89</td>\n",
       "      <td>47.596667</td>\n",
       "      <td>19.2</td>\n",
       "      <td>44.790000</td>\n",
       "      <td>19.79</td>\n",
       "      <td>44.730000</td>\n",
       "      <td>19.000000</td>\n",
       "      <td>45.566667</td>\n",
       "      <td>...</td>\n",
       "      <td>17.033333</td>\n",
       "      <td>45.53</td>\n",
       "      <td>6.600000</td>\n",
       "      <td>733.5</td>\n",
       "      <td>92.0</td>\n",
       "      <td>7.000000</td>\n",
       "      <td>63.000000</td>\n",
       "      <td>5.3</td>\n",
       "      <td>13.275433</td>\n",
       "      <td>13.275433</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2016-01-11 17:10:00</th>\n",
       "      <td>60</td>\n",
       "      <td>30</td>\n",
       "      <td>19.89</td>\n",
       "      <td>46.693333</td>\n",
       "      <td>19.2</td>\n",
       "      <td>44.722500</td>\n",
       "      <td>19.79</td>\n",
       "      <td>44.790000</td>\n",
       "      <td>19.000000</td>\n",
       "      <td>45.992500</td>\n",
       "      <td>...</td>\n",
       "      <td>17.066667</td>\n",
       "      <td>45.56</td>\n",
       "      <td>6.483333</td>\n",
       "      <td>733.6</td>\n",
       "      <td>92.0</td>\n",
       "      <td>6.666667</td>\n",
       "      <td>59.166667</td>\n",
       "      <td>5.2</td>\n",
       "      <td>18.606195</td>\n",
       "      <td>18.606195</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2016-01-11 17:20:00</th>\n",
       "      <td>50</td>\n",
       "      <td>30</td>\n",
       "      <td>19.89</td>\n",
       "      <td>46.300000</td>\n",
       "      <td>19.2</td>\n",
       "      <td>44.626667</td>\n",
       "      <td>19.79</td>\n",
       "      <td>44.933333</td>\n",
       "      <td>18.926667</td>\n",
       "      <td>45.890000</td>\n",
       "      <td>...</td>\n",
       "      <td>17.000000</td>\n",
       "      <td>45.50</td>\n",
       "      <td>6.366667</td>\n",
       "      <td>733.7</td>\n",
       "      <td>92.0</td>\n",
       "      <td>6.333333</td>\n",
       "      <td>55.333333</td>\n",
       "      <td>5.1</td>\n",
       "      <td>28.642668</td>\n",
       "      <td>28.642668</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2016-01-11 17:30:00</th>\n",
       "      <td>50</td>\n",
       "      <td>40</td>\n",
       "      <td>19.89</td>\n",
       "      <td>46.066667</td>\n",
       "      <td>19.2</td>\n",
       "      <td>44.590000</td>\n",
       "      <td>19.79</td>\n",
       "      <td>45.000000</td>\n",
       "      <td>18.890000</td>\n",
       "      <td>45.723333</td>\n",
       "      <td>...</td>\n",
       "      <td>17.000000</td>\n",
       "      <td>45.40</td>\n",
       "      <td>6.250000</td>\n",
       "      <td>733.8</td>\n",
       "      <td>92.0</td>\n",
       "      <td>6.000000</td>\n",
       "      <td>51.500000</td>\n",
       "      <td>5.0</td>\n",
       "      <td>45.410389</td>\n",
       "      <td>45.410389</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2016-01-11 17:40:00</th>\n",
       "      <td>60</td>\n",
       "      <td>40</td>\n",
       "      <td>19.89</td>\n",
       "      <td>46.333333</td>\n",
       "      <td>19.2</td>\n",
       "      <td>44.530000</td>\n",
       "      <td>19.79</td>\n",
       "      <td>45.000000</td>\n",
       "      <td>18.890000</td>\n",
       "      <td>45.530000</td>\n",
       "      <td>...</td>\n",
       "      <td>17.000000</td>\n",
       "      <td>45.40</td>\n",
       "      <td>6.133333</td>\n",
       "      <td>733.9</td>\n",
       "      <td>92.0</td>\n",
       "      <td>5.666667</td>\n",
       "      <td>47.666667</td>\n",
       "      <td>4.9</td>\n",
       "      <td>10.084097</td>\n",
       "      <td>10.084097</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>5 rows × 28 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "                     Appliances  lights     T1       RH_1    T2       RH_2  \\\n",
       "date                                                                         \n",
       "2016-01-11 17:00:00          60      30  19.89  47.596667  19.2  44.790000   \n",
       "2016-01-11 17:10:00          60      30  19.89  46.693333  19.2  44.722500   \n",
       "2016-01-11 17:20:00          50      30  19.89  46.300000  19.2  44.626667   \n",
       "2016-01-11 17:30:00          50      40  19.89  46.066667  19.2  44.590000   \n",
       "2016-01-11 17:40:00          60      40  19.89  46.333333  19.2  44.530000   \n",
       "\n",
       "                        T3       RH_3         T4       RH_4  ...         T9  \\\n",
       "date                                                         ...              \n",
       "2016-01-11 17:00:00  19.79  44.730000  19.000000  45.566667  ...  17.033333   \n",
       "2016-01-11 17:10:00  19.79  44.790000  19.000000  45.992500  ...  17.066667   \n",
       "2016-01-11 17:20:00  19.79  44.933333  18.926667  45.890000  ...  17.000000   \n",
       "2016-01-11 17:30:00  19.79  45.000000  18.890000  45.723333  ...  17.000000   \n",
       "2016-01-11 17:40:00  19.79  45.000000  18.890000  45.530000  ...  17.000000   \n",
       "\n",
       "                      RH_9     T_out  Press_mm_hg  RH_out  Windspeed  \\\n",
       "date                                                                   \n",
       "2016-01-11 17:00:00  45.53  6.600000        733.5    92.0   7.000000   \n",
       "2016-01-11 17:10:00  45.56  6.483333        733.6    92.0   6.666667   \n",
       "2016-01-11 17:20:00  45.50  6.366667        733.7    92.0   6.333333   \n",
       "2016-01-11 17:30:00  45.40  6.250000        733.8    92.0   6.000000   \n",
       "2016-01-11 17:40:00  45.40  6.133333        733.9    92.0   5.666667   \n",
       "\n",
       "                     Visibility  Tdewpoint        rv1        rv2  \n",
       "date                                                              \n",
       "2016-01-11 17:00:00   63.000000        5.3  13.275433  13.275433  \n",
       "2016-01-11 17:10:00   59.166667        5.2  18.606195  18.606195  \n",
       "2016-01-11 17:20:00   55.333333        5.1  28.642668  28.642668  \n",
       "2016-01-11 17:30:00   51.500000        5.0  45.410389  45.410389  \n",
       "2016-01-11 17:40:00   47.666667        4.9  10.084097  10.084097  \n",
       "\n",
       "[5 rows x 28 columns]"
      ]
     },
     "execution_count": 8,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "<class 'pandas.core.frame.DataFrame'>\n",
      "Index: 19735 entries, 2016-01-11 17:00:00 to 2016-05-27 18:00:00\n",
      "Data columns (total 28 columns):\n",
      " #   Column       Non-Null Count  Dtype  \n",
      "---  ------       --------------  -----  \n",
      " 0   Appliances   19735 non-null  int64  \n",
      " 1   lights       19735 non-null  int64  \n",
      " 2   T1           19735 non-null  float64\n",
      " 3   RH_1         19735 non-null  float64\n",
      " 4   T2           19735 non-null  float64\n",
      " 5   RH_2         19735 non-null  float64\n",
      " 6   T3           19735 non-null  float64\n",
      " 7   RH_3         19735 non-null  float64\n",
      " 8   T4           19735 non-null  float64\n",
      " 9   RH_4         19735 non-null  float64\n",
      " 10  T5           19735 non-null  float64\n",
      " 11  RH_5         19735 non-null  float64\n",
      " 12  T6           19735 non-null  float64\n",
      " 13  RH_6         19735 non-null  float64\n",
      " 14  T7           19735 non-null  float64\n",
      " 15  RH_7         19735 non-null  float64\n",
      " 16  T8           19735 non-null  float64\n",
      " 17  RH_8         19735 non-null  float64\n",
      " 18  T9           19735 non-null  float64\n",
      " 19  RH_9         19735 non-null  float64\n",
      " 20  T_out        19735 non-null  float64\n",
      " 21  Press_mm_hg  19735 non-null  float64\n",
      " 22  RH_out       19735 non-null  float64\n",
      " 23  Windspeed    19735 non-null  float64\n",
      " 24  Visibility   19735 non-null  float64\n",
      " 25  Tdewpoint    19735 non-null  float64\n",
      " 26  rv1          19735 non-null  float64\n",
      " 27  rv2          19735 non-null  float64\n",
      "dtypes: float64(26), int64(2)\n",
      "memory usage: 4.4+ MB\n"
     ]
    }
   ],
   "source": [
    "df.info()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "-------"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Quick EDA"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Index(['Appliances', 'lights', 'T1', 'RH_1', 'T2', 'RH_2', 'T3', 'RH_3', 'T4',\n",
       "       'RH_4', 'T5', 'RH_5', 'T6', 'RH_6', 'T7', 'RH_7', 'T8', 'RH_8', 'T9',\n",
       "       'RH_9', 'T_out', 'Press_mm_hg', 'RH_out', 'Windspeed', 'Visibility',\n",
       "       'Tdewpoint', 'rv1', 'rv2'],\n",
       "      dtype='object')"
      ]
     },
     "execution_count": 10,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df.columns"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 864x576 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "df['Windspeed'].plot(figsize=(12,8));\n",
    "plt.xticks(rotation=30);"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 864x648 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "df['Appliances'].plot(figsize=(12,9));\n",
    "plt.xticks(rotation=30);"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "-----"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Train Test Split"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "19735"
      ]
     },
     "execution_count": 18,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "len(df)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Let's imagine we want to predict just 24 hours into the future, we don't need 3 months of data for that, so let's save some training time and only select the last months data."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "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>Appliances</th>\n",
       "      <th>lights</th>\n",
       "      <th>T1</th>\n",
       "      <th>RH_1</th>\n",
       "      <th>T2</th>\n",
       "      <th>RH_2</th>\n",
       "      <th>T3</th>\n",
       "      <th>RH_3</th>\n",
       "      <th>T4</th>\n",
       "      <th>RH_4</th>\n",
       "      <th>...</th>\n",
       "      <th>T9</th>\n",
       "      <th>RH_9</th>\n",
       "      <th>T_out</th>\n",
       "      <th>Press_mm_hg</th>\n",
       "      <th>RH_out</th>\n",
       "      <th>Windspeed</th>\n",
       "      <th>Visibility</th>\n",
       "      <th>Tdewpoint</th>\n",
       "      <th>rv1</th>\n",
       "      <th>rv2</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",
       "      <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",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>2016-05-27 17:20:00</th>\n",
       "      <td>100</td>\n",
       "      <td>0</td>\n",
       "      <td>25.566667</td>\n",
       "      <td>46.560000</td>\n",
       "      <td>25.890000</td>\n",
       "      <td>42.025714</td>\n",
       "      <td>27.200000</td>\n",
       "      <td>41.163333</td>\n",
       "      <td>24.7</td>\n",
       "      <td>45.590000</td>\n",
       "      <td>...</td>\n",
       "      <td>23.2</td>\n",
       "      <td>46.7900</td>\n",
       "      <td>22.733333</td>\n",
       "      <td>755.2</td>\n",
       "      <td>55.666667</td>\n",
       "      <td>3.333333</td>\n",
       "      <td>23.666667</td>\n",
       "      <td>13.333333</td>\n",
       "      <td>43.096812</td>\n",
       "      <td>43.096812</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2016-05-27 17:30:00</th>\n",
       "      <td>90</td>\n",
       "      <td>0</td>\n",
       "      <td>25.500000</td>\n",
       "      <td>46.500000</td>\n",
       "      <td>25.754000</td>\n",
       "      <td>42.080000</td>\n",
       "      <td>27.133333</td>\n",
       "      <td>41.223333</td>\n",
       "      <td>24.7</td>\n",
       "      <td>45.590000</td>\n",
       "      <td>...</td>\n",
       "      <td>23.2</td>\n",
       "      <td>46.7900</td>\n",
       "      <td>22.600000</td>\n",
       "      <td>755.2</td>\n",
       "      <td>56.000000</td>\n",
       "      <td>3.500000</td>\n",
       "      <td>24.500000</td>\n",
       "      <td>13.300000</td>\n",
       "      <td>49.282940</td>\n",
       "      <td>49.282940</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2016-05-27 17:40:00</th>\n",
       "      <td>270</td>\n",
       "      <td>10</td>\n",
       "      <td>25.500000</td>\n",
       "      <td>46.596667</td>\n",
       "      <td>25.628571</td>\n",
       "      <td>42.768571</td>\n",
       "      <td>27.050000</td>\n",
       "      <td>41.690000</td>\n",
       "      <td>24.7</td>\n",
       "      <td>45.730000</td>\n",
       "      <td>...</td>\n",
       "      <td>23.2</td>\n",
       "      <td>46.7900</td>\n",
       "      <td>22.466667</td>\n",
       "      <td>755.2</td>\n",
       "      <td>56.333333</td>\n",
       "      <td>3.666667</td>\n",
       "      <td>25.333333</td>\n",
       "      <td>13.266667</td>\n",
       "      <td>29.199117</td>\n",
       "      <td>29.199117</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2016-05-27 17:50:00</th>\n",
       "      <td>420</td>\n",
       "      <td>10</td>\n",
       "      <td>25.500000</td>\n",
       "      <td>46.990000</td>\n",
       "      <td>25.414000</td>\n",
       "      <td>43.036000</td>\n",
       "      <td>26.890000</td>\n",
       "      <td>41.290000</td>\n",
       "      <td>24.7</td>\n",
       "      <td>45.790000</td>\n",
       "      <td>...</td>\n",
       "      <td>23.2</td>\n",
       "      <td>46.8175</td>\n",
       "      <td>22.333333</td>\n",
       "      <td>755.2</td>\n",
       "      <td>56.666667</td>\n",
       "      <td>3.833333</td>\n",
       "      <td>26.166667</td>\n",
       "      <td>13.233333</td>\n",
       "      <td>6.322784</td>\n",
       "      <td>6.322784</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2016-05-27 18:00:00</th>\n",
       "      <td>430</td>\n",
       "      <td>10</td>\n",
       "      <td>25.500000</td>\n",
       "      <td>46.600000</td>\n",
       "      <td>25.264286</td>\n",
       "      <td>42.971429</td>\n",
       "      <td>26.823333</td>\n",
       "      <td>41.156667</td>\n",
       "      <td>24.7</td>\n",
       "      <td>45.963333</td>\n",
       "      <td>...</td>\n",
       "      <td>23.2</td>\n",
       "      <td>46.8450</td>\n",
       "      <td>22.200000</td>\n",
       "      <td>755.2</td>\n",
       "      <td>57.000000</td>\n",
       "      <td>4.000000</td>\n",
       "      <td>27.000000</td>\n",
       "      <td>13.200000</td>\n",
       "      <td>34.118851</td>\n",
       "      <td>34.118851</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>5 rows × 28 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "                     Appliances  lights         T1       RH_1         T2  \\\n",
       "date                                                                       \n",
       "2016-05-27 17:20:00         100       0  25.566667  46.560000  25.890000   \n",
       "2016-05-27 17:30:00          90       0  25.500000  46.500000  25.754000   \n",
       "2016-05-27 17:40:00         270      10  25.500000  46.596667  25.628571   \n",
       "2016-05-27 17:50:00         420      10  25.500000  46.990000  25.414000   \n",
       "2016-05-27 18:00:00         430      10  25.500000  46.600000  25.264286   \n",
       "\n",
       "                          RH_2         T3       RH_3    T4       RH_4  ...  \\\n",
       "date                                                                   ...   \n",
       "2016-05-27 17:20:00  42.025714  27.200000  41.163333  24.7  45.590000  ...   \n",
       "2016-05-27 17:30:00  42.080000  27.133333  41.223333  24.7  45.590000  ...   \n",
       "2016-05-27 17:40:00  42.768571  27.050000  41.690000  24.7  45.730000  ...   \n",
       "2016-05-27 17:50:00  43.036000  26.890000  41.290000  24.7  45.790000  ...   \n",
       "2016-05-27 18:00:00  42.971429  26.823333  41.156667  24.7  45.963333  ...   \n",
       "\n",
       "                       T9     RH_9      T_out  Press_mm_hg     RH_out  \\\n",
       "date                                                                    \n",
       "2016-05-27 17:20:00  23.2  46.7900  22.733333        755.2  55.666667   \n",
       "2016-05-27 17:30:00  23.2  46.7900  22.600000        755.2  56.000000   \n",
       "2016-05-27 17:40:00  23.2  46.7900  22.466667        755.2  56.333333   \n",
       "2016-05-27 17:50:00  23.2  46.8175  22.333333        755.2  56.666667   \n",
       "2016-05-27 18:00:00  23.2  46.8450  22.200000        755.2  57.000000   \n",
       "\n",
       "                     Windspeed  Visibility  Tdewpoint        rv1        rv2  \n",
       "date                                                                         \n",
       "2016-05-27 17:20:00   3.333333   23.666667  13.333333  43.096812  43.096812  \n",
       "2016-05-27 17:30:00   3.500000   24.500000  13.300000  49.282940  49.282940  \n",
       "2016-05-27 17:40:00   3.666667   25.333333  13.266667  29.199117  29.199117  \n",
       "2016-05-27 17:50:00   3.833333   26.166667  13.233333   6.322784   6.322784  \n",
       "2016-05-27 18:00:00   4.000000   27.000000  13.200000  34.118851  34.118851  \n",
       "\n",
       "[5 rows x 28 columns]"
      ]
     },
     "execution_count": 19,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df.tail()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "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>Appliances</th>\n",
       "      <th>lights</th>\n",
       "      <th>T1</th>\n",
       "      <th>RH_1</th>\n",
       "      <th>T2</th>\n",
       "      <th>RH_2</th>\n",
       "      <th>T3</th>\n",
       "      <th>RH_3</th>\n",
       "      <th>T4</th>\n",
       "      <th>RH_4</th>\n",
       "      <th>...</th>\n",
       "      <th>T9</th>\n",
       "      <th>RH_9</th>\n",
       "      <th>T_out</th>\n",
       "      <th>Press_mm_hg</th>\n",
       "      <th>RH_out</th>\n",
       "      <th>Windspeed</th>\n",
       "      <th>Visibility</th>\n",
       "      <th>Tdewpoint</th>\n",
       "      <th>rv1</th>\n",
       "      <th>rv2</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",
       "      <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",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>2016-05-01 00:00:00</th>\n",
       "      <td>50</td>\n",
       "      <td>0</td>\n",
       "      <td>21.200000</td>\n",
       "      <td>38.200000</td>\n",
       "      <td>18.390000</td>\n",
       "      <td>41.400000</td>\n",
       "      <td>23.200000</td>\n",
       "      <td>36.400000</td>\n",
       "      <td>19.73</td>\n",
       "      <td>37.730000</td>\n",
       "      <td>...</td>\n",
       "      <td>19.166667</td>\n",
       "      <td>38.200000</td>\n",
       "      <td>4.300000</td>\n",
       "      <td>763.700000</td>\n",
       "      <td>86.000000</td>\n",
       "      <td>2.000000</td>\n",
       "      <td>40.000000</td>\n",
       "      <td>2.200000</td>\n",
       "      <td>15.310003</td>\n",
       "      <td>15.310003</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2016-05-01 00:10:00</th>\n",
       "      <td>60</td>\n",
       "      <td>0</td>\n",
       "      <td>21.150000</td>\n",
       "      <td>38.172500</td>\n",
       "      <td>18.390000</td>\n",
       "      <td>41.500000</td>\n",
       "      <td>23.133333</td>\n",
       "      <td>36.466667</td>\n",
       "      <td>19.70</td>\n",
       "      <td>37.663333</td>\n",
       "      <td>...</td>\n",
       "      <td>19.133333</td>\n",
       "      <td>38.290000</td>\n",
       "      <td>4.283333</td>\n",
       "      <td>763.716667</td>\n",
       "      <td>86.333333</td>\n",
       "      <td>2.166667</td>\n",
       "      <td>38.166667</td>\n",
       "      <td>2.216667</td>\n",
       "      <td>11.240067</td>\n",
       "      <td>11.240067</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2016-05-01 00:20:00</th>\n",
       "      <td>50</td>\n",
       "      <td>0</td>\n",
       "      <td>21.133333</td>\n",
       "      <td>38.090000</td>\n",
       "      <td>18.323333</td>\n",
       "      <td>41.500000</td>\n",
       "      <td>23.200000</td>\n",
       "      <td>36.500000</td>\n",
       "      <td>19.70</td>\n",
       "      <td>37.590000</td>\n",
       "      <td>...</td>\n",
       "      <td>19.133333</td>\n",
       "      <td>38.363333</td>\n",
       "      <td>4.266667</td>\n",
       "      <td>763.733333</td>\n",
       "      <td>86.666667</td>\n",
       "      <td>2.333333</td>\n",
       "      <td>36.333333</td>\n",
       "      <td>2.233333</td>\n",
       "      <td>21.808814</td>\n",
       "      <td>21.808814</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2016-05-01 00:30:00</th>\n",
       "      <td>50</td>\n",
       "      <td>0</td>\n",
       "      <td>21.100000</td>\n",
       "      <td>38.090000</td>\n",
       "      <td>18.290000</td>\n",
       "      <td>41.500000</td>\n",
       "      <td>23.200000</td>\n",
       "      <td>36.500000</td>\n",
       "      <td>19.70</td>\n",
       "      <td>37.500000</td>\n",
       "      <td>...</td>\n",
       "      <td>19.166667</td>\n",
       "      <td>38.500000</td>\n",
       "      <td>4.250000</td>\n",
       "      <td>763.750000</td>\n",
       "      <td>87.000000</td>\n",
       "      <td>2.500000</td>\n",
       "      <td>34.500000</td>\n",
       "      <td>2.250000</td>\n",
       "      <td>47.000534</td>\n",
       "      <td>47.000534</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2016-05-01 00:40:00</th>\n",
       "      <td>60</td>\n",
       "      <td>0</td>\n",
       "      <td>21.100000</td>\n",
       "      <td>38.030000</td>\n",
       "      <td>18.290000</td>\n",
       "      <td>41.560000</td>\n",
       "      <td>23.200000</td>\n",
       "      <td>36.500000</td>\n",
       "      <td>19.70</td>\n",
       "      <td>37.500000</td>\n",
       "      <td>...</td>\n",
       "      <td>19.166667</td>\n",
       "      <td>38.633333</td>\n",
       "      <td>4.233333</td>\n",
       "      <td>763.766667</td>\n",
       "      <td>87.333333</td>\n",
       "      <td>2.666667</td>\n",
       "      <td>32.666667</td>\n",
       "      <td>2.266667</td>\n",
       "      <td>8.059441</td>\n",
       "      <td>8.059441</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2016-05-27 17:20:00</th>\n",
       "      <td>100</td>\n",
       "      <td>0</td>\n",
       "      <td>25.566667</td>\n",
       "      <td>46.560000</td>\n",
       "      <td>25.890000</td>\n",
       "      <td>42.025714</td>\n",
       "      <td>27.200000</td>\n",
       "      <td>41.163333</td>\n",
       "      <td>24.70</td>\n",
       "      <td>45.590000</td>\n",
       "      <td>...</td>\n",
       "      <td>23.200000</td>\n",
       "      <td>46.790000</td>\n",
       "      <td>22.733333</td>\n",
       "      <td>755.200000</td>\n",
       "      <td>55.666667</td>\n",
       "      <td>3.333333</td>\n",
       "      <td>23.666667</td>\n",
       "      <td>13.333333</td>\n",
       "      <td>43.096812</td>\n",
       "      <td>43.096812</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2016-05-27 17:30:00</th>\n",
       "      <td>90</td>\n",
       "      <td>0</td>\n",
       "      <td>25.500000</td>\n",
       "      <td>46.500000</td>\n",
       "      <td>25.754000</td>\n",
       "      <td>42.080000</td>\n",
       "      <td>27.133333</td>\n",
       "      <td>41.223333</td>\n",
       "      <td>24.70</td>\n",
       "      <td>45.590000</td>\n",
       "      <td>...</td>\n",
       "      <td>23.200000</td>\n",
       "      <td>46.790000</td>\n",
       "      <td>22.600000</td>\n",
       "      <td>755.200000</td>\n",
       "      <td>56.000000</td>\n",
       "      <td>3.500000</td>\n",
       "      <td>24.500000</td>\n",
       "      <td>13.300000</td>\n",
       "      <td>49.282940</td>\n",
       "      <td>49.282940</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2016-05-27 17:40:00</th>\n",
       "      <td>270</td>\n",
       "      <td>10</td>\n",
       "      <td>25.500000</td>\n",
       "      <td>46.596667</td>\n",
       "      <td>25.628571</td>\n",
       "      <td>42.768571</td>\n",
       "      <td>27.050000</td>\n",
       "      <td>41.690000</td>\n",
       "      <td>24.70</td>\n",
       "      <td>45.730000</td>\n",
       "      <td>...</td>\n",
       "      <td>23.200000</td>\n",
       "      <td>46.790000</td>\n",
       "      <td>22.466667</td>\n",
       "      <td>755.200000</td>\n",
       "      <td>56.333333</td>\n",
       "      <td>3.666667</td>\n",
       "      <td>25.333333</td>\n",
       "      <td>13.266667</td>\n",
       "      <td>29.199117</td>\n",
       "      <td>29.199117</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2016-05-27 17:50:00</th>\n",
       "      <td>420</td>\n",
       "      <td>10</td>\n",
       "      <td>25.500000</td>\n",
       "      <td>46.990000</td>\n",
       "      <td>25.414000</td>\n",
       "      <td>43.036000</td>\n",
       "      <td>26.890000</td>\n",
       "      <td>41.290000</td>\n",
       "      <td>24.70</td>\n",
       "      <td>45.790000</td>\n",
       "      <td>...</td>\n",
       "      <td>23.200000</td>\n",
       "      <td>46.817500</td>\n",
       "      <td>22.333333</td>\n",
       "      <td>755.200000</td>\n",
       "      <td>56.666667</td>\n",
       "      <td>3.833333</td>\n",
       "      <td>26.166667</td>\n",
       "      <td>13.233333</td>\n",
       "      <td>6.322784</td>\n",
       "      <td>6.322784</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2016-05-27 18:00:00</th>\n",
       "      <td>430</td>\n",
       "      <td>10</td>\n",
       "      <td>25.500000</td>\n",
       "      <td>46.600000</td>\n",
       "      <td>25.264286</td>\n",
       "      <td>42.971429</td>\n",
       "      <td>26.823333</td>\n",
       "      <td>41.156667</td>\n",
       "      <td>24.70</td>\n",
       "      <td>45.963333</td>\n",
       "      <td>...</td>\n",
       "      <td>23.200000</td>\n",
       "      <td>46.845000</td>\n",
       "      <td>22.200000</td>\n",
       "      <td>755.200000</td>\n",
       "      <td>57.000000</td>\n",
       "      <td>4.000000</td>\n",
       "      <td>27.000000</td>\n",
       "      <td>13.200000</td>\n",
       "      <td>34.118851</td>\n",
       "      <td>34.118851</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>3853 rows × 28 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "                     Appliances  lights         T1       RH_1         T2  \\\n",
       "date                                                                       \n",
       "2016-05-01 00:00:00          50       0  21.200000  38.200000  18.390000   \n",
       "2016-05-01 00:10:00          60       0  21.150000  38.172500  18.390000   \n",
       "2016-05-01 00:20:00          50       0  21.133333  38.090000  18.323333   \n",
       "2016-05-01 00:30:00          50       0  21.100000  38.090000  18.290000   \n",
       "2016-05-01 00:40:00          60       0  21.100000  38.030000  18.290000   \n",
       "...                         ...     ...        ...        ...        ...   \n",
       "2016-05-27 17:20:00         100       0  25.566667  46.560000  25.890000   \n",
       "2016-05-27 17:30:00          90       0  25.500000  46.500000  25.754000   \n",
       "2016-05-27 17:40:00         270      10  25.500000  46.596667  25.628571   \n",
       "2016-05-27 17:50:00         420      10  25.500000  46.990000  25.414000   \n",
       "2016-05-27 18:00:00         430      10  25.500000  46.600000  25.264286   \n",
       "\n",
       "                          RH_2         T3       RH_3     T4       RH_4  ...  \\\n",
       "date                                                                    ...   \n",
       "2016-05-01 00:00:00  41.400000  23.200000  36.400000  19.73  37.730000  ...   \n",
       "2016-05-01 00:10:00  41.500000  23.133333  36.466667  19.70  37.663333  ...   \n",
       "2016-05-01 00:20:00  41.500000  23.200000  36.500000  19.70  37.590000  ...   \n",
       "2016-05-01 00:30:00  41.500000  23.200000  36.500000  19.70  37.500000  ...   \n",
       "2016-05-01 00:40:00  41.560000  23.200000  36.500000  19.70  37.500000  ...   \n",
       "...                        ...        ...        ...    ...        ...  ...   \n",
       "2016-05-27 17:20:00  42.025714  27.200000  41.163333  24.70  45.590000  ...   \n",
       "2016-05-27 17:30:00  42.080000  27.133333  41.223333  24.70  45.590000  ...   \n",
       "2016-05-27 17:40:00  42.768571  27.050000  41.690000  24.70  45.730000  ...   \n",
       "2016-05-27 17:50:00  43.036000  26.890000  41.290000  24.70  45.790000  ...   \n",
       "2016-05-27 18:00:00  42.971429  26.823333  41.156667  24.70  45.963333  ...   \n",
       "\n",
       "                            T9       RH_9      T_out  Press_mm_hg     RH_out  \\\n",
       "date                                                                           \n",
       "2016-05-01 00:00:00  19.166667  38.200000   4.300000   763.700000  86.000000   \n",
       "2016-05-01 00:10:00  19.133333  38.290000   4.283333   763.716667  86.333333   \n",
       "2016-05-01 00:20:00  19.133333  38.363333   4.266667   763.733333  86.666667   \n",
       "2016-05-01 00:30:00  19.166667  38.500000   4.250000   763.750000  87.000000   \n",
       "2016-05-01 00:40:00  19.166667  38.633333   4.233333   763.766667  87.333333   \n",
       "...                        ...        ...        ...          ...        ...   \n",
       "2016-05-27 17:20:00  23.200000  46.790000  22.733333   755.200000  55.666667   \n",
       "2016-05-27 17:30:00  23.200000  46.790000  22.600000   755.200000  56.000000   \n",
       "2016-05-27 17:40:00  23.200000  46.790000  22.466667   755.200000  56.333333   \n",
       "2016-05-27 17:50:00  23.200000  46.817500  22.333333   755.200000  56.666667   \n",
       "2016-05-27 18:00:00  23.200000  46.845000  22.200000   755.200000  57.000000   \n",
       "\n",
       "                     Windspeed  Visibility  Tdewpoint        rv1        rv2  \n",
       "date                                                                         \n",
       "2016-05-01 00:00:00   2.000000   40.000000   2.200000  15.310003  15.310003  \n",
       "2016-05-01 00:10:00   2.166667   38.166667   2.216667  11.240067  11.240067  \n",
       "2016-05-01 00:20:00   2.333333   36.333333   2.233333  21.808814  21.808814  \n",
       "2016-05-01 00:30:00   2.500000   34.500000   2.250000  47.000534  47.000534  \n",
       "2016-05-01 00:40:00   2.666667   32.666667   2.266667   8.059441   8.059441  \n",
       "...                        ...         ...        ...        ...        ...  \n",
       "2016-05-27 17:20:00   3.333333   23.666667  13.333333  43.096812  43.096812  \n",
       "2016-05-27 17:30:00   3.500000   24.500000  13.300000  49.282940  49.282940  \n",
       "2016-05-27 17:40:00   3.666667   25.333333  13.266667  29.199117  29.199117  \n",
       "2016-05-27 17:50:00   3.833333   26.166667  13.233333   6.322784   6.322784  \n",
       "2016-05-27 18:00:00   4.000000   27.000000  13.200000  34.118851  34.118851  \n",
       "\n",
       "[3853 rows x 28 columns]"
      ]
     },
     "execution_count": 21,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df.loc['2016-05-01':]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {},
   "outputs": [],
   "source": [
    "df = df.loc['2016-05-01':]"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "If we look at the data such as `T1`, we can see there are so many precision for each values.\n",
    "\n",
    "Let's also round off the data, to one decimal point precision, otherwise this may cause issues with our network (we will also normalize the data anyways, so this level of precision isn't useful to us)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {},
   "outputs": [],
   "source": [
    "df = df.round(2)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "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>Appliances</th>\n",
       "      <th>lights</th>\n",
       "      <th>T1</th>\n",
       "      <th>RH_1</th>\n",
       "      <th>T2</th>\n",
       "      <th>RH_2</th>\n",
       "      <th>T3</th>\n",
       "      <th>RH_3</th>\n",
       "      <th>T4</th>\n",
       "      <th>RH_4</th>\n",
       "      <th>...</th>\n",
       "      <th>T9</th>\n",
       "      <th>RH_9</th>\n",
       "      <th>T_out</th>\n",
       "      <th>Press_mm_hg</th>\n",
       "      <th>RH_out</th>\n",
       "      <th>Windspeed</th>\n",
       "      <th>Visibility</th>\n",
       "      <th>Tdewpoint</th>\n",
       "      <th>rv1</th>\n",
       "      <th>rv2</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",
       "      <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",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>2016-05-01 00:00:00</th>\n",
       "      <td>50</td>\n",
       "      <td>0</td>\n",
       "      <td>21.20</td>\n",
       "      <td>38.20</td>\n",
       "      <td>18.39</td>\n",
       "      <td>41.4</td>\n",
       "      <td>23.20</td>\n",
       "      <td>36.40</td>\n",
       "      <td>19.73</td>\n",
       "      <td>37.73</td>\n",
       "      <td>...</td>\n",
       "      <td>19.17</td>\n",
       "      <td>38.20</td>\n",
       "      <td>4.30</td>\n",
       "      <td>763.70</td>\n",
       "      <td>86.00</td>\n",
       "      <td>2.00</td>\n",
       "      <td>40.00</td>\n",
       "      <td>2.20</td>\n",
       "      <td>15.31</td>\n",
       "      <td>15.31</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2016-05-01 00:10:00</th>\n",
       "      <td>60</td>\n",
       "      <td>0</td>\n",
       "      <td>21.15</td>\n",
       "      <td>38.17</td>\n",
       "      <td>18.39</td>\n",
       "      <td>41.5</td>\n",
       "      <td>23.13</td>\n",
       "      <td>36.47</td>\n",
       "      <td>19.70</td>\n",
       "      <td>37.66</td>\n",
       "      <td>...</td>\n",
       "      <td>19.13</td>\n",
       "      <td>38.29</td>\n",
       "      <td>4.28</td>\n",
       "      <td>763.72</td>\n",
       "      <td>86.33</td>\n",
       "      <td>2.17</td>\n",
       "      <td>38.17</td>\n",
       "      <td>2.22</td>\n",
       "      <td>11.24</td>\n",
       "      <td>11.24</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>2 rows × 28 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "                     Appliances  lights     T1   RH_1     T2  RH_2     T3  \\\n",
       "date                                                                        \n",
       "2016-05-01 00:00:00          50       0  21.20  38.20  18.39  41.4  23.20   \n",
       "2016-05-01 00:10:00          60       0  21.15  38.17  18.39  41.5  23.13   \n",
       "\n",
       "                      RH_3     T4   RH_4  ...     T9   RH_9  T_out  \\\n",
       "date                                      ...                        \n",
       "2016-05-01 00:00:00  36.40  19.73  37.73  ...  19.17  38.20   4.30   \n",
       "2016-05-01 00:10:00  36.47  19.70  37.66  ...  19.13  38.29   4.28   \n",
       "\n",
       "                     Press_mm_hg  RH_out  Windspeed  Visibility  Tdewpoint  \\\n",
       "date                                                                         \n",
       "2016-05-01 00:00:00       763.70   86.00       2.00       40.00       2.20   \n",
       "2016-05-01 00:10:00       763.72   86.33       2.17       38.17       2.22   \n",
       "\n",
       "                       rv1    rv2  \n",
       "date                               \n",
       "2016-05-01 00:00:00  15.31  15.31  \n",
       "2016-05-01 00:10:00  11.24  11.24  \n",
       "\n",
       "[2 rows x 28 columns]"
      ]
     },
     "execution_count": 26,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df.head(2)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Check how batch size for sequence will be looked like.\n",
    "\n",
    "As data are recorded per 10 mintues, we need to calculate how to set up for each batch size."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "3853"
      ]
     },
     "execution_count": 27,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "len(df)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "144.0"
      ]
     },
     "execution_count": 28,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# How many rows per day? We know its every 10 min\n",
    "# total minutes per day / frequency of 10 minutes => total records per day\n",
    "24 * 60 / 10"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "metadata": {},
   "outputs": [],
   "source": [
    "test_days = 2 # we want to use 2 days of data as testing data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 32,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "288"
      ]
     },
     "execution_count": 32,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "test_index_starting_point = test_days * 144\n",
    "test_index_starting_point"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 33,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Actual Splitting\n",
    "# Take note of - minus sign in indexing, we need to leave 288 rows as test data\n",
    "train = df.iloc[:-test_index_starting_point]\n",
    "test = df.iloc[-test_index_starting_point: ]\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 34,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "((3565, 28), (288, 28))"
      ]
     },
     "execution_count": 34,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "train.shape, test.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 35,
   "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>Appliances</th>\n",
       "      <th>lights</th>\n",
       "      <th>T1</th>\n",
       "      <th>RH_1</th>\n",
       "      <th>T2</th>\n",
       "      <th>RH_2</th>\n",
       "      <th>T3</th>\n",
       "      <th>RH_3</th>\n",
       "      <th>T4</th>\n",
       "      <th>RH_4</th>\n",
       "      <th>...</th>\n",
       "      <th>T9</th>\n",
       "      <th>RH_9</th>\n",
       "      <th>T_out</th>\n",
       "      <th>Press_mm_hg</th>\n",
       "      <th>RH_out</th>\n",
       "      <th>Windspeed</th>\n",
       "      <th>Visibility</th>\n",
       "      <th>Tdewpoint</th>\n",
       "      <th>rv1</th>\n",
       "      <th>rv2</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",
       "      <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",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>2016-05-25 17:50:00</th>\n",
       "      <td>90</td>\n",
       "      <td>0</td>\n",
       "      <td>24.5</td>\n",
       "      <td>37.63</td>\n",
       "      <td>24.03</td>\n",
       "      <td>34.43</td>\n",
       "      <td>25.1</td>\n",
       "      <td>36.99</td>\n",
       "      <td>24.29</td>\n",
       "      <td>37.93</td>\n",
       "      <td>...</td>\n",
       "      <td>22.0</td>\n",
       "      <td>37.36</td>\n",
       "      <td>16.42</td>\n",
       "      <td>756.12</td>\n",
       "      <td>55.17</td>\n",
       "      <td>1.83</td>\n",
       "      <td>37.83</td>\n",
       "      <td>7.38</td>\n",
       "      <td>39.36</td>\n",
       "      <td>39.36</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2016-05-25 18:00:00</th>\n",
       "      <td>100</td>\n",
       "      <td>0</td>\n",
       "      <td>24.5</td>\n",
       "      <td>38.00</td>\n",
       "      <td>24.00</td>\n",
       "      <td>34.40</td>\n",
       "      <td>25.1</td>\n",
       "      <td>36.73</td>\n",
       "      <td>24.29</td>\n",
       "      <td>37.86</td>\n",
       "      <td>...</td>\n",
       "      <td>22.0</td>\n",
       "      <td>37.36</td>\n",
       "      <td>16.50</td>\n",
       "      <td>756.10</td>\n",
       "      <td>56.00</td>\n",
       "      <td>2.00</td>\n",
       "      <td>40.00</td>\n",
       "      <td>7.70</td>\n",
       "      <td>38.63</td>\n",
       "      <td>38.63</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>2 rows × 28 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "                     Appliances  lights    T1   RH_1     T2   RH_2    T3  \\\n",
       "date                                                                       \n",
       "2016-05-25 17:50:00          90       0  24.5  37.63  24.03  34.43  25.1   \n",
       "2016-05-25 18:00:00         100       0  24.5  38.00  24.00  34.40  25.1   \n",
       "\n",
       "                      RH_3     T4   RH_4  ...    T9   RH_9  T_out  \\\n",
       "date                                      ...                       \n",
       "2016-05-25 17:50:00  36.99  24.29  37.93  ...  22.0  37.36  16.42   \n",
       "2016-05-25 18:00:00  36.73  24.29  37.86  ...  22.0  37.36  16.50   \n",
       "\n",
       "                     Press_mm_hg  RH_out  Windspeed  Visibility  Tdewpoint  \\\n",
       "date                                                                         \n",
       "2016-05-25 17:50:00       756.12   55.17       1.83       37.83       7.38   \n",
       "2016-05-25 18:00:00       756.10   56.00       2.00       40.00       7.70   \n",
       "\n",
       "                       rv1    rv2  \n",
       "date                               \n",
       "2016-05-25 17:50:00  39.36  39.36  \n",
       "2016-05-25 18:00:00  38.63  38.63  \n",
       "\n",
       "[2 rows x 28 columns]"
      ]
     },
     "execution_count": 35,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "train.tail(2)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 36,
   "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>Appliances</th>\n",
       "      <th>lights</th>\n",
       "      <th>T1</th>\n",
       "      <th>RH_1</th>\n",
       "      <th>T2</th>\n",
       "      <th>RH_2</th>\n",
       "      <th>T3</th>\n",
       "      <th>RH_3</th>\n",
       "      <th>T4</th>\n",
       "      <th>RH_4</th>\n",
       "      <th>...</th>\n",
       "      <th>T9</th>\n",
       "      <th>RH_9</th>\n",
       "      <th>T_out</th>\n",
       "      <th>Press_mm_hg</th>\n",
       "      <th>RH_out</th>\n",
       "      <th>Windspeed</th>\n",
       "      <th>Visibility</th>\n",
       "      <th>Tdewpoint</th>\n",
       "      <th>rv1</th>\n",
       "      <th>rv2</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",
       "      <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",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>2016-05-25 18:10:00</th>\n",
       "      <td>170</td>\n",
       "      <td>0</td>\n",
       "      <td>24.5</td>\n",
       "      <td>37.86</td>\n",
       "      <td>24.00</td>\n",
       "      <td>34.27</td>\n",
       "      <td>25.00</td>\n",
       "      <td>36.70</td>\n",
       "      <td>24.29</td>\n",
       "      <td>37.79</td>\n",
       "      <td>...</td>\n",
       "      <td>22.0</td>\n",
       "      <td>37.23</td>\n",
       "      <td>16.48</td>\n",
       "      <td>756.1</td>\n",
       "      <td>55.83</td>\n",
       "      <td>2.0</td>\n",
       "      <td>38.17</td>\n",
       "      <td>7.62</td>\n",
       "      <td>38.14</td>\n",
       "      <td>38.14</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2016-05-25 18:20:00</th>\n",
       "      <td>310</td>\n",
       "      <td>0</td>\n",
       "      <td>24.5</td>\n",
       "      <td>37.30</td>\n",
       "      <td>23.86</td>\n",
       "      <td>34.33</td>\n",
       "      <td>24.94</td>\n",
       "      <td>36.67</td>\n",
       "      <td>24.29</td>\n",
       "      <td>37.79</td>\n",
       "      <td>...</td>\n",
       "      <td>22.0</td>\n",
       "      <td>37.36</td>\n",
       "      <td>16.47</td>\n",
       "      <td>756.1</td>\n",
       "      <td>55.67</td>\n",
       "      <td>2.0</td>\n",
       "      <td>36.33</td>\n",
       "      <td>7.53</td>\n",
       "      <td>31.16</td>\n",
       "      <td>31.16</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>2 rows × 28 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "                     Appliances  lights    T1   RH_1     T2   RH_2     T3  \\\n",
       "date                                                                        \n",
       "2016-05-25 18:10:00         170       0  24.5  37.86  24.00  34.27  25.00   \n",
       "2016-05-25 18:20:00         310       0  24.5  37.30  23.86  34.33  24.94   \n",
       "\n",
       "                      RH_3     T4   RH_4  ...    T9   RH_9  T_out  \\\n",
       "date                                      ...                       \n",
       "2016-05-25 18:10:00  36.70  24.29  37.79  ...  22.0  37.23  16.48   \n",
       "2016-05-25 18:20:00  36.67  24.29  37.79  ...  22.0  37.36  16.47   \n",
       "\n",
       "                     Press_mm_hg  RH_out  Windspeed  Visibility  Tdewpoint  \\\n",
       "date                                                                         \n",
       "2016-05-25 18:10:00        756.1   55.83        2.0       38.17       7.62   \n",
       "2016-05-25 18:20:00        756.1   55.67        2.0       36.33       7.53   \n",
       "\n",
       "                       rv1    rv2  \n",
       "date                               \n",
       "2016-05-25 18:10:00  38.14  38.14  \n",
       "2016-05-25 18:20:00  31.16  31.16  \n",
       "\n",
       "[2 rows x 28 columns]"
      ]
     },
     "execution_count": 36,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "test.head(2)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "---------"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Scale Data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 37,
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.preprocessing import MinMaxScaler"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 38,
   "metadata": {},
   "outputs": [],
   "source": [
    "scaler = MinMaxScaler()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 39,
   "metadata": {},
   "outputs": [],
   "source": [
    "scaled_train = scaler.fit_transform(train)\n",
    "scaled_test = scaler.transform(test)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "------"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Time Series Generator\n",
    "\n",
    "This class takes in a sequence of data-points gathered at\n",
    "equal intervals, along with time series parameters such as\n",
    "stride, length of history, etc., to produce batches for\n",
    "training/validation.\n",
    "\n",
    "#### Arguments\n",
    "    data: Indexable generator (such as list or Numpy array)\n",
    "        containing consecutive data points (timesteps).\n",
    "        The data should be at 2D, and axis 0 is expected\n",
    "        to be the time dimension.\n",
    "    targets: Targets corresponding to timesteps in `data`.\n",
    "        It should have same length as `data`.\n",
    "    length: Length of the output sequences (in number of timesteps).\n",
    "    sampling_rate: Period between successive individual timesteps\n",
    "        within sequences. For rate `r`, timesteps\n",
    "        `data[i]`, `data[i-r]`, ... `data[i - length]`\n",
    "        are used for create a sample sequence.\n",
    "    stride: Period between successive output sequences.\n",
    "        For stride `s`, consecutive output samples would\n",
    "        be centered around `data[i]`, `data[i+s]`, `data[i+2*s]`, etc.\n",
    "    start_index: Data points earlier than `start_index` will not be used\n",
    "        in the output sequences. This is useful to reserve part of the\n",
    "        data for test or validation.\n",
    "    end_index: Data points later than `end_index` will not be used\n",
    "        in the output sequences. This is useful to reserve part of the\n",
    "        data for test or validation.\n",
    "    shuffle: Whether to shuffle output samples,\n",
    "        or instead draw them in chronological order.\n",
    "    reverse: Boolean: if `true`, timesteps in each output sample will be\n",
    "        in reverse chronological order.\n",
    "    batch_size: Number of timeseries samples in each batch\n",
    "        (except maybe the last one)."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 44,
   "metadata": {},
   "outputs": [],
   "source": [
    "from tensorflow.keras.preprocessing.sequence import TimeseriesGenerator"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**NOTE: edit the length so that it makes sense for your time series!**"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 45,
   "metadata": {},
   "outputs": [],
   "source": [
    "length = 144 # for one day full daily cycle\n",
    "batch_size = 1"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 47,
   "metadata": {},
   "outputs": [],
   "source": [
    "generator = TimeseriesGenerator(data=scaled_train, targets=scaled_train,\n",
    "                                                   length=length, batch_size=1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 50,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Given the Array: \n",
      "[0.03896104 0.         0.13798978 ... 0.14319527 0.75185111 0.75185111]\n",
      "\n",
      "Predict this y: \n",
      " [[0.03896104 0.         0.30834753 0.29439421 0.16038492 0.49182278\n",
      "  0.0140056  0.36627907 0.24142857 0.24364791 0.12650602 0.36276002\n",
      "  0.12       0.28205572 0.06169297 0.15759185 0.34582624 0.39585974\n",
      "  0.09259259 0.39649608 0.18852459 0.96052632 0.59210526 0.1\n",
      "  0.58333333 0.13609467 0.4576746  0.4576746 ]]\n"
     ]
    }
   ],
   "source": [
    "# What does the first batch look like?\n",
    "X,y = generator[0]\n",
    "\n",
    "print(f'Given the Array: \\n{X.flatten()}')\n",
    "print(f'\\nPredict this y: \\n {y}')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "-------"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Model Building"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 49,
   "metadata": {},
   "outputs": [],
   "source": [
    "from tensorflow.keras.models import Sequential\n",
    "from tensorflow.keras.layers import Dense, LSTM"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Reflect the mulitvariate features for number of features "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 51,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(3565, 28)"
      ]
     },
     "execution_count": 51,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "scaled_train.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 53,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "28"
      ]
     },
     "execution_count": 53,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "n_features = scaled_train.shape[1]\n",
    "n_features"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 56,
   "metadata": {},
   "outputs": [],
   "source": [
    "model = Sequential()\n",
    "\n",
    "# Simple RNN Layer\n",
    "model.add(LSTM(100, input_shape=(length, n_features)))\n",
    "\n",
    "# Final Prediction (one neuron per feature)\n",
    "model.add(Dense(n_features))\n",
    "          \n",
    "model.compile(optimizer='adam', loss='mse')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 57,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Model: \"sequential\"\n",
      "_________________________________________________________________\n",
      "Layer (type)                 Output Shape              Param #   \n",
      "=================================================================\n",
      "lstm (LSTM)                  (None, 100)               51600     \n",
      "_________________________________________________________________\n",
      "dense (Dense)                (None, 28)                2828      \n",
      "=================================================================\n",
      "Total params: 54,428\n",
      "Trainable params: 54,428\n",
      "Non-trainable params: 0\n",
      "_________________________________________________________________\n"
     ]
    }
   ],
   "source": [
    "model.summary()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "-----"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# EarlyStopping and Validation Generator"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 59,
   "metadata": {},
   "outputs": [],
   "source": [
    "from tensorflow.keras.callbacks import EarlyStopping"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 60,
   "metadata": {},
   "outputs": [],
   "source": [
    "early_stopping = EarlyStopping(monitor='val_loss', patience=2, verbose=2)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 61,
   "metadata": {},
   "outputs": [],
   "source": [
    "validation_generator = TimeseriesGenerator(data=scaled_test, targets=scaled_test,\n",
    "                                                                      length=length, batch_size=batch_size)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "-----"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Model Training"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### As we are now dealing with Multivariate Timeseries and there are a lot of features, we need to consider lower number of epochs to train on."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 62,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 1/2\n",
      "3421/3421 [==============================] - 152s 44ms/step - loss: 0.0116 - val_loss: 0.0112\n",
      "Epoch 2/2\n",
      "3421/3421 [==============================] - 157s 46ms/step - loss: 0.0079 - val_loss: 0.0092\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "<tensorflow.python.keras.callbacks.History at 0x27ef391dc0>"
      ]
     },
     "execution_count": 62,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "model.fit(generator, \n",
    "             epochs=2,\n",
    "             validation_data = validation_generator,\n",
    "             callbacks=[early_stopping],\n",
    "             verbose=1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 63,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "losses = pd.DataFrame(model.history.history)\n",
    "losses.plot();"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "----------"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Evaluate on Test Data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 64,
   "metadata": {},
   "outputs": [],
   "source": [
    "test_predictions = []\n",
    "\n",
    "first_evaluation_batch = scaled_train[-length: ]\n",
    "current_batch = first_evaluation_batch.reshape((1, length, n_features))\n",
    "\n",
    "for i in range(len(test)):\n",
    "    current_prediction = model.predict(current_batch)[0]\n",
    "    \n",
    "    test_predictions.append(current_prediction)\n",
    "    \n",
    "    current_batch = np.append(current_batch[:, 1:, :], [[current_prediction]], axis=1)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Inverse Transformations and Compare"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 65,
   "metadata": {},
   "outputs": [],
   "source": [
    "true_predictions = scaler.inverse_transform(test_predictions)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 66,
   "metadata": {},
   "outputs": [],
   "source": [
    "true_predictions_df = pd.DataFrame(data=true_predictions, columns=test.columns)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 67,
   "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>Appliances</th>\n",
       "      <th>lights</th>\n",
       "      <th>T1</th>\n",
       "      <th>RH_1</th>\n",
       "      <th>T2</th>\n",
       "      <th>RH_2</th>\n",
       "      <th>T3</th>\n",
       "      <th>RH_3</th>\n",
       "      <th>T4</th>\n",
       "      <th>RH_4</th>\n",
       "      <th>...</th>\n",
       "      <th>T9</th>\n",
       "      <th>RH_9</th>\n",
       "      <th>T_out</th>\n",
       "      <th>Press_mm_hg</th>\n",
       "      <th>RH_out</th>\n",
       "      <th>Windspeed</th>\n",
       "      <th>Visibility</th>\n",
       "      <th>Tdewpoint</th>\n",
       "      <th>rv1</th>\n",
       "      <th>rv2</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>105.513212</td>\n",
       "      <td>3.320100</td>\n",
       "      <td>24.434543</td>\n",
       "      <td>38.031104</td>\n",
       "      <td>24.277211</td>\n",
       "      <td>35.276308</td>\n",
       "      <td>25.079605</td>\n",
       "      <td>37.023892</td>\n",
       "      <td>23.952222</td>\n",
       "      <td>37.022506</td>\n",
       "      <td>...</td>\n",
       "      <td>22.072506</td>\n",
       "      <td>38.023030</td>\n",
       "      <td>16.795420</td>\n",
       "      <td>756.066685</td>\n",
       "      <td>57.700998</td>\n",
       "      <td>2.230769</td>\n",
       "      <td>36.865144</td>\n",
       "      <td>7.569594</td>\n",
       "      <td>23.760460</td>\n",
       "      <td>25.647860</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>111.244663</td>\n",
       "      <td>4.528213</td>\n",
       "      <td>24.435531</td>\n",
       "      <td>38.172667</td>\n",
       "      <td>24.231856</td>\n",
       "      <td>35.853338</td>\n",
       "      <td>25.067250</td>\n",
       "      <td>37.070755</td>\n",
       "      <td>23.903068</td>\n",
       "      <td>37.137094</td>\n",
       "      <td>...</td>\n",
       "      <td>22.105541</td>\n",
       "      <td>38.337015</td>\n",
       "      <td>17.017254</td>\n",
       "      <td>756.027694</td>\n",
       "      <td>58.103538</td>\n",
       "      <td>2.426337</td>\n",
       "      <td>35.476775</td>\n",
       "      <td>7.815580</td>\n",
       "      <td>24.569094</td>\n",
       "      <td>26.158112</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>113.741887</td>\n",
       "      <td>5.332637</td>\n",
       "      <td>24.428873</td>\n",
       "      <td>38.248952</td>\n",
       "      <td>24.185040</td>\n",
       "      <td>36.153335</td>\n",
       "      <td>25.045987</td>\n",
       "      <td>37.103238</td>\n",
       "      <td>23.873696</td>\n",
       "      <td>37.235559</td>\n",
       "      <td>...</td>\n",
       "      <td>22.141369</td>\n",
       "      <td>38.639513</td>\n",
       "      <td>17.058367</td>\n",
       "      <td>755.921104</td>\n",
       "      <td>58.251036</td>\n",
       "      <td>2.572959</td>\n",
       "      <td>33.787146</td>\n",
       "      <td>7.942222</td>\n",
       "      <td>25.304431</td>\n",
       "      <td>26.955495</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>115.238806</td>\n",
       "      <td>5.907305</td>\n",
       "      <td>24.416213</td>\n",
       "      <td>38.305367</td>\n",
       "      <td>24.134080</td>\n",
       "      <td>36.391396</td>\n",
       "      <td>25.012960</td>\n",
       "      <td>37.126821</td>\n",
       "      <td>23.844445</td>\n",
       "      <td>37.307714</td>\n",
       "      <td>...</td>\n",
       "      <td>22.167315</td>\n",
       "      <td>38.912531</td>\n",
       "      <td>17.087300</td>\n",
       "      <td>755.845158</td>\n",
       "      <td>58.390923</td>\n",
       "      <td>2.697640</td>\n",
       "      <td>31.992881</td>\n",
       "      <td>8.011880</td>\n",
       "      <td>25.983915</td>\n",
       "      <td>27.645032</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>115.925140</td>\n",
       "      <td>6.326018</td>\n",
       "      <td>24.399831</td>\n",
       "      <td>38.343268</td>\n",
       "      <td>24.077480</td>\n",
       "      <td>36.598237</td>\n",
       "      <td>24.970189</td>\n",
       "      <td>37.140067</td>\n",
       "      <td>23.810039</td>\n",
       "      <td>37.360640</td>\n",
       "      <td>...</td>\n",
       "      <td>22.183287</td>\n",
       "      <td>39.152409</td>\n",
       "      <td>17.109250</td>\n",
       "      <td>755.805311</td>\n",
       "      <td>58.549426</td>\n",
       "      <td>2.817962</td>\n",
       "      <td>30.234244</td>\n",
       "      <td>8.048265</td>\n",
       "      <td>26.617117</td>\n",
       "      <td>28.266177</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>5 rows × 28 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "   Appliances    lights         T1       RH_1         T2       RH_2  \\\n",
       "0  105.513212  3.320100  24.434543  38.031104  24.277211  35.276308   \n",
       "1  111.244663  4.528213  24.435531  38.172667  24.231856  35.853338   \n",
       "2  113.741887  5.332637  24.428873  38.248952  24.185040  36.153335   \n",
       "3  115.238806  5.907305  24.416213  38.305367  24.134080  36.391396   \n",
       "4  115.925140  6.326018  24.399831  38.343268  24.077480  36.598237   \n",
       "\n",
       "          T3       RH_3         T4       RH_4  ...         T9       RH_9  \\\n",
       "0  25.079605  37.023892  23.952222  37.022506  ...  22.072506  38.023030   \n",
       "1  25.067250  37.070755  23.903068  37.137094  ...  22.105541  38.337015   \n",
       "2  25.045987  37.103238  23.873696  37.235559  ...  22.141369  38.639513   \n",
       "3  25.012960  37.126821  23.844445  37.307714  ...  22.167315  38.912531   \n",
       "4  24.970189  37.140067  23.810039  37.360640  ...  22.183287  39.152409   \n",
       "\n",
       "       T_out  Press_mm_hg     RH_out  Windspeed  Visibility  Tdewpoint  \\\n",
       "0  16.795420   756.066685  57.700998   2.230769   36.865144   7.569594   \n",
       "1  17.017254   756.027694  58.103538   2.426337   35.476775   7.815580   \n",
       "2  17.058367   755.921104  58.251036   2.572959   33.787146   7.942222   \n",
       "3  17.087300   755.845158  58.390923   2.697640   31.992881   8.011880   \n",
       "4  17.109250   755.805311  58.549426   2.817962   30.234244   8.048265   \n",
       "\n",
       "         rv1        rv2  \n",
       "0  23.760460  25.647860  \n",
       "1  24.569094  26.158112  \n",
       "2  25.304431  26.955495  \n",
       "3  25.983915  27.645032  \n",
       "4  26.617117  28.266177  \n",
       "\n",
       "[5 rows x 28 columns]"
      ]
     },
     "execution_count": 67,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "true_predictions_df.head()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Evaluation"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "+ If we want to evaluate our model predictions, we can compare RMSE on features that we consider important.\n",
    "+ Or if we want the RMSE as a whole for every features, we can cosider weight average\n",
    "+ However we need to take note that every features scale are different, we might want to use scaled version values to compare againt each other."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Save Model"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 68,
   "metadata": {},
   "outputs": [],
   "source": [
    "from tensorflow.keras.models import load_model"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 69,
   "metadata": {},
   "outputs": [],
   "source": [
    "model.save('../Models/multivariate.h5')"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "venv-datascience",
   "language": "python",
   "name": "venv-datascience"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.8.5"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 4
}