LSTM per hour 0.83% .ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "id": "ea3afff5",
   "metadata": {},
   "source": [
    "using the by Hour data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "id": "f48586f5",
   "metadata": {},
   "outputs": [],
   "source": [
    "import pandas as pd, matplotlib as plt , seaborn as sns, numpy as np, tensorflow as tf\n",
    "from sklearn.model_selection import train_test_split\n",
    "from sklearn.preprocessing import MinMaxScaler"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "id": "3021f092",
   "metadata": {},
   "outputs": [],
   "source": [
    "df = pd.read_csv('content/preprocessed_train.csv').query('building_id==122 & meter==0').drop(columns= ['Unnamed: 0','precip_depth_1_hr','cloud_coverage', 'site_id', 'square_feet', 'building_id','primary_use', 'meter'] )"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "id": "95ba50ad",
   "metadata": {},
   "outputs": [],
   "source": [
    "df[\"timestamp\"] = pd.to_datetime(df[\"timestamp\"])\n",
    "df.set_index('timestamp', inplace=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "id": "e4bdbd66",
   "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>meter_reading</th>\n",
       "      <th>air_temperature</th>\n",
       "      <th>dew_temperature</th>\n",
       "      <th>sea_level_pressure</th>\n",
       "      <th>wind_direction</th>\n",
       "      <th>wind_speed</th>\n",
       "      <th>day</th>\n",
       "      <th>month</th>\n",
       "      <th>hour</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>timestamp</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>2016-01-01 00:00:00</th>\n",
       "      <td>134.3</td>\n",
       "      <td>3.8</td>\n",
       "      <td>2.4</td>\n",
       "      <td>1020.9</td>\n",
       "      <td>240.0</td>\n",
       "      <td>3.1</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2016-01-01 01:00:00</th>\n",
       "      <td>267.1</td>\n",
       "      <td>3.7</td>\n",
       "      <td>2.4</td>\n",
       "      <td>1021.6</td>\n",
       "      <td>230.0</td>\n",
       "      <td>2.6</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2016-01-01 02:00:00</th>\n",
       "      <td>268.7</td>\n",
       "      <td>2.6</td>\n",
       "      <td>1.9</td>\n",
       "      <td>1021.9</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>2</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2016-01-01 03:00:00</th>\n",
       "      <td>267.7</td>\n",
       "      <td>2.0</td>\n",
       "      <td>1.2</td>\n",
       "      <td>1022.3</td>\n",
       "      <td>170.0</td>\n",
       "      <td>1.5</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>3</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2016-01-01 04:00:00</th>\n",
       "      <td>267.8</td>\n",
       "      <td>2.3</td>\n",
       "      <td>1.8</td>\n",
       "      <td>1022.7</td>\n",
       "      <td>110.0</td>\n",
       "      <td>1.5</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>4</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                     meter_reading  air_temperature  dew_temperature  \\\n",
       "timestamp                                                              \n",
       "2016-01-01 00:00:00          134.3              3.8              2.4   \n",
       "2016-01-01 01:00:00          267.1              3.7              2.4   \n",
       "2016-01-01 02:00:00          268.7              2.6              1.9   \n",
       "2016-01-01 03:00:00          267.7              2.0              1.2   \n",
       "2016-01-01 04:00:00          267.8              2.3              1.8   \n",
       "\n",
       "                     sea_level_pressure  wind_direction  wind_speed  day  \\\n",
       "timestamp                                                                  \n",
       "2016-01-01 00:00:00              1020.9           240.0         3.1    1   \n",
       "2016-01-01 01:00:00              1021.6           230.0         2.6    1   \n",
       "2016-01-01 02:00:00              1021.9             0.0         0.0    1   \n",
       "2016-01-01 03:00:00              1022.3           170.0         1.5    1   \n",
       "2016-01-01 04:00:00              1022.7           110.0         1.5    1   \n",
       "\n",
       "                     month  hour  \n",
       "timestamp                         \n",
       "2016-01-01 00:00:00      1     0  \n",
       "2016-01-01 01:00:00      1     1  \n",
       "2016-01-01 02:00:00      1     2  \n",
       "2016-01-01 03:00:00      1     3  \n",
       "2016-01-01 04:00:00      1     4  "
      ]
     },
     "execution_count": 4,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 62,
   "id": "36e2986e",
   "metadata": {
    "scrolled": false
   },
   "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>meter_reading</th>\n",
       "      <th>air_temperature</th>\n",
       "      <th>dew_temperature</th>\n",
       "      <th>sea_level_pressure</th>\n",
       "      <th>wind_direction</th>\n",
       "      <th>wind_speed</th>\n",
       "      <th>day</th>\n",
       "      <th>month</th>\n",
       "      <th>hour</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>timestamp</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>2016-01-01 00:00:00</th>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.209239</td>\n",
       "      <td>0.381757</td>\n",
       "      <td>0.659280</td>\n",
       "      <td>0.666667</td>\n",
       "      <td>0.167568</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2016-01-01 01:00:00</th>\n",
       "      <td>0.364635</td>\n",
       "      <td>0.206522</td>\n",
       "      <td>0.381757</td>\n",
       "      <td>0.668975</td>\n",
       "      <td>0.638889</td>\n",
       "      <td>0.140541</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2016-01-01 02:00:00</th>\n",
       "      <td>0.369028</td>\n",
       "      <td>0.176630</td>\n",
       "      <td>0.364865</td>\n",
       "      <td>0.673130</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>2</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2016-01-01 03:00:00</th>\n",
       "      <td>0.366282</td>\n",
       "      <td>0.160326</td>\n",
       "      <td>0.341216</td>\n",
       "      <td>0.678670</td>\n",
       "      <td>0.472222</td>\n",
       "      <td>0.081081</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>3</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2016-01-01 04:00:00</th>\n",
       "      <td>0.366557</td>\n",
       "      <td>0.168478</td>\n",
       "      <td>0.361486</td>\n",
       "      <td>0.684211</td>\n",
       "      <td>0.305556</td>\n",
       "      <td>0.081081</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>4</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                     meter_reading  air_temperature  dew_temperature  \\\n",
       "timestamp                                                              \n",
       "2016-01-01 00:00:00       0.000000         0.209239         0.381757   \n",
       "2016-01-01 01:00:00       0.364635         0.206522         0.381757   \n",
       "2016-01-01 02:00:00       0.369028         0.176630         0.364865   \n",
       "2016-01-01 03:00:00       0.366282         0.160326         0.341216   \n",
       "2016-01-01 04:00:00       0.366557         0.168478         0.361486   \n",
       "\n",
       "                     sea_level_pressure  wind_direction  wind_speed  day  \\\n",
       "timestamp                                                                  \n",
       "2016-01-01 00:00:00            0.659280        0.666667    0.167568    1   \n",
       "2016-01-01 01:00:00            0.668975        0.638889    0.140541    1   \n",
       "2016-01-01 02:00:00            0.673130        0.000000    0.000000    1   \n",
       "2016-01-01 03:00:00            0.678670        0.472222    0.081081    1   \n",
       "2016-01-01 04:00:00            0.684211        0.305556    0.081081    1   \n",
       "\n",
       "                     month  hour  \n",
       "timestamp                         \n",
       "2016-01-01 00:00:00      1     0  \n",
       "2016-01-01 01:00:00      1     1  \n",
       "2016-01-01 02:00:00      1     2  \n",
       "2016-01-01 03:00:00      1     3  \n",
       "2016-01-01 04:00:00      1     4  "
      ]
     },
     "execution_count": 62,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "scaler=MinMaxScaler()\n",
    "# only scaling the numerical values not day/month\n",
    "scaled_data=scaler.fit_transform(df.drop(columns=['day','month','hour']))\n",
    "\n",
    "scaled_df= pd.concat( [ pd.DataFrame(scaled_data, index=df.index , columns=df.columns[:-3]) , df.iloc[:,-3:]  ] \n",
    "                     , axis=1 )\n",
    "scaled_df.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 116,
   "id": "d780d62a",
   "metadata": {},
   "outputs": [],
   "source": [
    "filled_data= scaled_df.fillna(method='ffill', axis=1 )"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 117,
   "id": "ff8e3c3e",
   "metadata": {},
   "outputs": [],
   "source": [
    "x_train,x_val,y_train,y_val= train_test_split( filled_data.drop(columns='meter_reading') ,\n",
    "                                                    filled_data.meter_reading,\n",
    "                                                    test_size=0.2,\n",
    "                                                    shuffle=False )\n",
    "assert len(x_train) == len(y_train)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 118,
   "id": "4dfde94b",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "timestamp\n",
       "2016-10-19 20:00:00    0.335255\n",
       "2016-10-19 21:00:00    0.286107\n",
       "2016-10-19 22:00:00    0.276496\n",
       "2016-10-19 23:00:00    0.266337\n",
       "2016-10-20 00:00:00    0.266612\n",
       "                         ...   \n",
       "2016-12-31 19:00:00    0.292696\n",
       "2016-12-31 20:00:00    0.285283\n",
       "2016-12-31 21:00:00    0.280066\n",
       "2016-12-31 22:00:00    0.275947\n",
       "2016-12-31 23:00:00    0.278418\n",
       "Name: meter_reading, Length: 1753, dtype: float64"
      ]
     },
     "execution_count": 118,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "y_val"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 119,
   "id": "02914607",
   "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>air_temperature</th>\n",
       "      <th>dew_temperature</th>\n",
       "      <th>sea_level_pressure</th>\n",
       "      <th>wind_direction</th>\n",
       "      <th>wind_speed</th>\n",
       "      <th>day</th>\n",
       "      <th>month</th>\n",
       "      <th>hour</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>timestamp</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>2016-10-19 20:00:00</th>\n",
       "      <td>0.413043</td>\n",
       "      <td>0.520270</td>\n",
       "      <td>0.668975</td>\n",
       "      <td>0.888889</td>\n",
       "      <td>0.140541</td>\n",
       "      <td>19.0</td>\n",
       "      <td>10.0</td>\n",
       "      <td>20.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2016-10-19 21:00:00</th>\n",
       "      <td>0.402174</td>\n",
       "      <td>0.530405</td>\n",
       "      <td>0.670360</td>\n",
       "      <td>0.861111</td>\n",
       "      <td>0.140541</td>\n",
       "      <td>19.0</td>\n",
       "      <td>10.0</td>\n",
       "      <td>21.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2016-10-19 22:00:00</th>\n",
       "      <td>0.391304</td>\n",
       "      <td>0.523649</td>\n",
       "      <td>0.671745</td>\n",
       "      <td>0.833333</td>\n",
       "      <td>0.167568</td>\n",
       "      <td>19.0</td>\n",
       "      <td>10.0</td>\n",
       "      <td>22.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2016-10-19 23:00:00</th>\n",
       "      <td>0.366848</td>\n",
       "      <td>0.523649</td>\n",
       "      <td>0.670360</td>\n",
       "      <td>0.861111</td>\n",
       "      <td>0.113514</td>\n",
       "      <td>19.0</td>\n",
       "      <td>10.0</td>\n",
       "      <td>23.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2016-10-20 00:00:00</th>\n",
       "      <td>0.334239</td>\n",
       "      <td>0.513514</td>\n",
       "      <td>0.667590</td>\n",
       "      <td>0.888889</td>\n",
       "      <td>0.113514</td>\n",
       "      <td>20.0</td>\n",
       "      <td>10.0</td>\n",
       "      <td>0.0</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",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2016-12-31 19:00:00</th>\n",
       "      <td>0.326087</td>\n",
       "      <td>0.520270</td>\n",
       "      <td>0.750693</td>\n",
       "      <td>0.611111</td>\n",
       "      <td>0.194595</td>\n",
       "      <td>31.0</td>\n",
       "      <td>12.0</td>\n",
       "      <td>19.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2016-12-31 20:00:00</th>\n",
       "      <td>0.301630</td>\n",
       "      <td>0.506757</td>\n",
       "      <td>0.742382</td>\n",
       "      <td>0.611111</td>\n",
       "      <td>0.221622</td>\n",
       "      <td>31.0</td>\n",
       "      <td>12.0</td>\n",
       "      <td>20.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2016-12-31 21:00:00</th>\n",
       "      <td>0.293478</td>\n",
       "      <td>0.496622</td>\n",
       "      <td>0.732687</td>\n",
       "      <td>0.611111</td>\n",
       "      <td>0.248649</td>\n",
       "      <td>31.0</td>\n",
       "      <td>12.0</td>\n",
       "      <td>21.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2016-12-31 22:00:00</th>\n",
       "      <td>0.293478</td>\n",
       "      <td>0.510135</td>\n",
       "      <td>0.721607</td>\n",
       "      <td>0.527778</td>\n",
       "      <td>0.167568</td>\n",
       "      <td>31.0</td>\n",
       "      <td>12.0</td>\n",
       "      <td>22.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2016-12-31 23:00:00</th>\n",
       "      <td>0.309783</td>\n",
       "      <td>0.520270</td>\n",
       "      <td>0.711911</td>\n",
       "      <td>0.583333</td>\n",
       "      <td>0.308108</td>\n",
       "      <td>31.0</td>\n",
       "      <td>12.0</td>\n",
       "      <td>23.0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>1753 rows × 8 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "                     air_temperature  dew_temperature  sea_level_pressure  \\\n",
       "timestamp                                                                   \n",
       "2016-10-19 20:00:00         0.413043         0.520270            0.668975   \n",
       "2016-10-19 21:00:00         0.402174         0.530405            0.670360   \n",
       "2016-10-19 22:00:00         0.391304         0.523649            0.671745   \n",
       "2016-10-19 23:00:00         0.366848         0.523649            0.670360   \n",
       "2016-10-20 00:00:00         0.334239         0.513514            0.667590   \n",
       "...                              ...              ...                 ...   \n",
       "2016-12-31 19:00:00         0.326087         0.520270            0.750693   \n",
       "2016-12-31 20:00:00         0.301630         0.506757            0.742382   \n",
       "2016-12-31 21:00:00         0.293478         0.496622            0.732687   \n",
       "2016-12-31 22:00:00         0.293478         0.510135            0.721607   \n",
       "2016-12-31 23:00:00         0.309783         0.520270            0.711911   \n",
       "\n",
       "                     wind_direction  wind_speed   day  month  hour  \n",
       "timestamp                                                           \n",
       "2016-10-19 20:00:00        0.888889    0.140541  19.0   10.0  20.0  \n",
       "2016-10-19 21:00:00        0.861111    0.140541  19.0   10.0  21.0  \n",
       "2016-10-19 22:00:00        0.833333    0.167568  19.0   10.0  22.0  \n",
       "2016-10-19 23:00:00        0.861111    0.113514  19.0   10.0  23.0  \n",
       "2016-10-20 00:00:00        0.888889    0.113514  20.0   10.0   0.0  \n",
       "...                             ...         ...   ...    ...   ...  \n",
       "2016-12-31 19:00:00        0.611111    0.194595  31.0   12.0  19.0  \n",
       "2016-12-31 20:00:00        0.611111    0.221622  31.0   12.0  20.0  \n",
       "2016-12-31 21:00:00        0.611111    0.248649  31.0   12.0  21.0  \n",
       "2016-12-31 22:00:00        0.527778    0.167568  31.0   12.0  22.0  \n",
       "2016-12-31 23:00:00        0.583333    0.308108  31.0   12.0  23.0  \n",
       "\n",
       "[1753 rows x 8 columns]"
      ]
     },
     "execution_count": 119,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "x_val"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 120,
   "id": "a9de4994",
   "metadata": {},
   "outputs": [],
   "source": [
    "# creating csv files of the data\n",
    "list_of_df= [x_train , y_train,\n",
    "            x_val , y_val]\n",
    "\n",
    "list_of_df_names= ['x_train_h' , 'y_train_h',\n",
    "                    'x_val_h' , 'y_val_h']\n",
    "\n",
    "for i, frame in enumerate(list_of_df):\n",
    "    frame.to_csv( f'content/{list_of_df_names[i]}.csv')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 121,
   "id": "e1f7f164",
   "metadata": {},
   "outputs": [],
   "source": [
    "list_of_df=  ['x_train_h' , 'y_train_h',\n",
    "                    'x_val_h' , 'y_val_h']\n",
    "\n",
    "for i, frame in enumerate(list_of_df):\n",
    "    list_of_df[i]= pd.read_csv( f'content/{list_of_df[i]}.csv')\n",
    "    \n",
    "x_train , y_train, x_val , y_val = list_of_df\n",
    "            "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 122,
   "id": "0ab933de",
   "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>timestamp</th>\n",
       "      <th>air_temperature</th>\n",
       "      <th>dew_temperature</th>\n",
       "      <th>sea_level_pressure</th>\n",
       "      <th>wind_direction</th>\n",
       "      <th>wind_speed</th>\n",
       "      <th>day</th>\n",
       "      <th>month</th>\n",
       "      <th>hour</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>2016-01-01 00:00:00</td>\n",
       "      <td>0.209239</td>\n",
       "      <td>0.381757</td>\n",
       "      <td>0.659280</td>\n",
       "      <td>0.666667</td>\n",
       "      <td>0.167568</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>2016-01-01 01:00:00</td>\n",
       "      <td>0.206522</td>\n",
       "      <td>0.381757</td>\n",
       "      <td>0.668975</td>\n",
       "      <td>0.638889</td>\n",
       "      <td>0.140541</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>2016-01-01 02:00:00</td>\n",
       "      <td>0.176630</td>\n",
       "      <td>0.364865</td>\n",
       "      <td>0.673130</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>2.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>2016-01-01 03:00:00</td>\n",
       "      <td>0.160326</td>\n",
       "      <td>0.341216</td>\n",
       "      <td>0.678670</td>\n",
       "      <td>0.472222</td>\n",
       "      <td>0.081081</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>3.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>2016-01-01 04:00:00</td>\n",
       "      <td>0.168478</td>\n",
       "      <td>0.361486</td>\n",
       "      <td>0.684211</td>\n",
       "      <td>0.305556</td>\n",
       "      <td>0.081081</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>4.0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "             timestamp  air_temperature  dew_temperature  sea_level_pressure  \\\n",
       "0  2016-01-01 00:00:00         0.209239         0.381757            0.659280   \n",
       "1  2016-01-01 01:00:00         0.206522         0.381757            0.668975   \n",
       "2  2016-01-01 02:00:00         0.176630         0.364865            0.673130   \n",
       "3  2016-01-01 03:00:00         0.160326         0.341216            0.678670   \n",
       "4  2016-01-01 04:00:00         0.168478         0.361486            0.684211   \n",
       "\n",
       "   wind_direction  wind_speed  day  month  hour  \n",
       "0        0.666667    0.167568  1.0    1.0   0.0  \n",
       "1        0.638889    0.140541  1.0    1.0   1.0  \n",
       "2        0.000000    0.000000  1.0    1.0   2.0  \n",
       "3        0.472222    0.081081  1.0    1.0   3.0  \n",
       "4        0.305556    0.081081  1.0    1.0   4.0  "
      ]
     },
     "execution_count": 122,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "x_train.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 123,
   "id": "a259b7de",
   "metadata": {},
   "outputs": [],
   "source": [
    "train_gen= tf.keras.preprocessing.sequence.TimeseriesGenerator(np.array( x_train.drop(columns='timestamp')), \n",
    "                                                               np.array( y_train.drop(columns='timestamp')) , \n",
    "                                                                length=6 ,sampling_rate=1, \n",
    "                                                                stride=1, batch_size=32\n",
    "                                                                )\n",
    "\n",
    "val_gen= tf.keras.preprocessing.sequence.TimeseriesGenerator(  np.array(x_val.drop(columns='timestamp') ), \n",
    "                                                               np.array(y_val.drop(columns='timestamp') ) , \n",
    "                                                                length=6 ,sampling_rate=1, \n",
    "                                                                stride=1, batch_size=32\n",
    "                                                                )\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 137,
   "id": "9543fc6d",
   "metadata": {},
   "outputs": [],
   "source": [
    "model= tf.keras.Sequential([tf.keras.layers.LSTM(32 , activation='relu' ,return_sequences=False),\n",
    "        tf.keras.layers.Dropout(0.3),\n",
    "        tf.keras.layers.Dense(1) ] )"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 138,
   "id": "132fc4f0",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 1/30\n",
      "214/219 [============================>.] - ETA: 0s - loss: 0.3619WARNING:tensorflow:From C:\\ProgramData\\Anaconda3\\lib\\site-packages\\tensorflow\\python\\training\\tracking\\tracking.py:111: Model.state_updates (from tensorflow.python.keras.engine.training) is deprecated and will be removed in a future version.\n",
      "Instructions for updating:\n",
      "This property should not be used in TensorFlow 2.0, as updates are applied automatically.\n",
      "WARNING:tensorflow:From C:\\ProgramData\\Anaconda3\\lib\\site-packages\\tensorflow\\python\\training\\tracking\\tracking.py:111: Layer.updates (from tensorflow.python.keras.engine.base_layer) is deprecated and will be removed in a future version.\n",
      "Instructions for updating:\n",
      "This property should not be used in TensorFlow 2.0, as updates are applied automatically.\n",
      "INFO:tensorflow:Assets written to: models\\LSTM_per_hour_model\\assets\n",
      "219/219 [==============================] - 9s 41ms/step - loss: 0.3544 - val_loss: 0.0131\n",
      "Epoch 2/30\n",
      "206/219 [===========================>..] - ETA: 0s - loss: 0.0393INFO:tensorflow:Assets written to: models\\LSTM_per_hour_model\\assets\n",
      "219/219 [==============================] - 9s 41ms/step - loss: 0.0382 - val_loss: 0.0115\n",
      "Epoch 3/30\n",
      "212/219 [============================>.] - ETA: 0s - loss: 0.0257INFO:tensorflow:Assets written to: models\\LSTM_per_hour_model\\assets\n",
      "219/219 [==============================] - 9s 43ms/step - loss: 0.0252 - val_loss: 0.0110\n",
      "Epoch 4/30\n",
      "219/219 [==============================] - 1s 5ms/step - loss: 0.0193 - val_loss: 0.0110\n",
      "Epoch 5/30\n",
      "210/219 [===========================>..] - ETA: 0s - loss: 0.0173INFO:tensorflow:Assets written to: models\\LSTM_per_hour_model\\assets\n",
      "219/219 [==============================] - 10s 43ms/step - loss: 0.0172 - val_loss: 0.0104\n",
      "Epoch 6/30\n",
      "214/219 [============================>.] - ETA: 0s - loss: 0.0152INFO:tensorflow:Assets written to: models\\LSTM_per_hour_model\\assets\n",
      "219/219 [==============================] - 8s 38ms/step - loss: 0.0151 - val_loss: 0.0091\n",
      "Epoch 7/30\n",
      "219/219 [==============================] - 1s 5ms/step - loss: 0.0145 - val_loss: 0.0097\n",
      "Epoch 8/30\n",
      "219/219 [==============================] - 1s 5ms/step - loss: 0.0140 - val_loss: 0.0100\n",
      "Epoch 9/30\n",
      "219/219 [==============================] - ETA: 0s - loss: 0.0132INFO:tensorflow:Assets written to: models\\LSTM_per_hour_model\\assets\n",
      "219/219 [==============================] - 9s 42ms/step - loss: 0.0132 - val_loss: 0.0088\n",
      "Epoch 10/30\n",
      "210/219 [===========================>..] - ETA: 0s - loss: 0.0131INFO:tensorflow:Assets written to: models\\LSTM_per_hour_model\\assets\n",
      "219/219 [==============================] - 9s 42ms/step - loss: 0.0129 - val_loss: 0.0087\n",
      "Epoch 11/30\n",
      "210/219 [===========================>..] - ETA: 0s - loss: 0.0123INFO:tensorflow:Assets written to: models\\LSTM_per_hour_model\\assets\n",
      "219/219 [==============================] - 8s 37ms/step - loss: 0.0121 - val_loss: 0.0079\n",
      "Epoch 12/30\n",
      "219/219 [==============================] - 1s 5ms/step - loss: 0.0118 - val_loss: 0.0089\n",
      "Epoch 13/30\n",
      "219/219 [==============================] - 1s 5ms/step - loss: 0.0116 - val_loss: 0.0086\n",
      "Epoch 14/30\n",
      "219/219 [==============================] - 1s 5ms/step - loss: 0.0113 - val_loss: 0.0090\n",
      "Epoch 15/30\n",
      "219/219 [==============================] - 1s 4ms/step - loss: 0.0112 - val_loss: 0.0099\n",
      "Epoch 16/30\n",
      "219/219 [==============================] - 1s 5ms/step - loss: 0.0113 - val_loss: 0.0087\n",
      "Epoch 17/30\n",
      "219/219 [==============================] - 1s 5ms/step - loss: 0.0110 - val_loss: 0.0093\n",
      "Epoch 18/30\n",
      "219/219 [==============================] - 1s 5ms/step - loss: 0.0108 - val_loss: 0.0103\n",
      "Epoch 19/30\n",
      "219/219 [==============================] - 1s 5ms/step - loss: 0.0108 - val_loss: 0.0080\n",
      "Epoch 20/30\n",
      "219/219 [==============================] - 1s 5ms/step - loss: 0.0105 - val_loss: 0.0089\n",
      "Epoch 21/30\n",
      "219/219 [==============================] - 1s 5ms/step - loss: 0.0102 - val_loss: 0.0080\n",
      "Epoch 22/30\n",
      "219/219 [==============================] - 1s 5ms/step - loss: 0.0105 - val_loss: 0.0087\n",
      "Epoch 23/30\n",
      "215/219 [============================>.] - ETA: 0s - loss: 0.0104INFO:tensorflow:Assets written to: models\\LSTM_per_hour_model\\assets\n",
      "219/219 [==============================] - 9s 43ms/step - loss: 0.0103 - val_loss: 0.0079\n",
      "Epoch 24/30\n",
      "219/219 [==============================] - 1s 5ms/step - loss: 0.0101 - val_loss: 0.0086\n",
      "Epoch 25/30\n",
      "219/219 [==============================] - 1s 5ms/step - loss: 0.0100 - val_loss: 0.0083\n",
      "Epoch 26/30\n",
      "219/219 [==============================] - 1s 5ms/step - loss: 0.0100 - val_loss: 0.0085\n",
      "Epoch 27/30\n",
      "219/219 [==============================] - 1s 5ms/step - loss: 0.0095 - val_loss: 0.0084\n",
      "Epoch 28/30\n",
      "219/219 [==============================] - 1s 5ms/step - loss: 0.0098 - val_loss: 0.0084\n",
      "Epoch 29/30\n",
      "219/219 [==============================] - 1s 4ms/step - loss: 0.0098 - val_loss: 0.0084\n",
      "Epoch 30/30\n",
      "219/219 [==============================] - ETA: 0s - loss: 0.0097INFO:tensorflow:Assets written to: models\\LSTM_per_hour_model\\assets\n",
      "219/219 [==============================] - 9s 42ms/step - loss: 0.0097 - val_loss: 0.0077\n"
     ]
    }
   ],
   "source": [
    "model.compile(loss='mse', optimizer= tf.keras.optimizers.Adam(0.001))\n",
    "\n",
    "cb= tf.keras.callbacks.ModelCheckpoint('models/LSTM_per_hour_model', monitor='val_loss', verbose=0, save_best_only=True)\n",
    "# Fitting the model\n",
    "history=model.fit(train_gen, \n",
    "          validation_data= val_gen,\n",
    "          epochs=30 ,\n",
    "          callbacks=[cb],\n",
    "          shuffle=False)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "64b22513",
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3 (ipykernel)",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.8.8"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}