bike_sharing_regression.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## **Predicting Bike Sharing Demand with Regression Models** \n",
    "---\n",
    "By Aditya Nugraha"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "<!-- ![image: https://www.freepik.com/vectors/ride-bicycle](assets/5243.jpg) -->\n",
    "<img src=\"assets/5243.jpg\" alt=\"image: https://www.freepik.com/vectors/ride-bicycle\"/>\n",
    "\n",
    "*Image Source: https://www.freepik.com/vectors/ride-bicycle*"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### **Contents**\n",
    "---\n",
    "\n",
    "1. Business Problem Understanding\n",
    "2. Data Understanding\n",
    "3. EDA & Feature Engineering\n",
    "4. Modeling\n",
    "5. Conclusion\n",
    "6. Recommendation\n",
    "7. Save Model"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### **Business Problem Understanding**\n",
    "---"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**Context**\n",
    "\n",
    "Sistem penyewaan sepeda adalah sarana penyewaan sepeda dimana proses mendapatkan keanggotaan, sewa, dan pengembalian sepeda secara mandiri melalui jaringan lokasi seperti kios di seluruh kota. Dengan menggunakan sistem ini, orang dapat menyewa sepeda dari satu lokasi dan mengembalikannya ke lokasi lain sesuai kebutuhan. Menurut data yang di rilis pada situs [Statista.com](https://www.statista.com/chart/14542/bike-sharing-programs-worldwide/) Hingga pertengahan tahun 2018, ada sekitar 1600 program penyewaan sepeda dengan 18,2 juta sepeda di lebih dari 1000 kota.\n",
    "\n",
    "<img src=\"assets/14542.jpeg\" alt=\"image: https://www.statista.com/chart/14542/bike-sharing-programs-worldwide/\" width=\"600\"/>\n",
    "\n",
    "*Image Source: https://www.statista.com/chart/14542/bike-sharing-programs-worldwide/*"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**Problem Statement**\n",
    "\n",
    "Terlepas dari efek positifnya pada lalu lintas, masalah lingkungan dan kesehatan, sistem penyewaan sepeda sangat menarik karena data yang mereka hasilkan dapat digunakan untuk tujuan penelitian. Waktu keberangkatan dan kedatangan, lokasi dan durasi penggunaan dicatat dalam sistem ini. Data tersebut dapat digunakan untuk memprediksi kapan waktu, kondisi, dan tempat yang tepat bagi masyarakat untuk melakukan penyewaan sepeda, sekaligus menjadi masalah tersendiri bagi penyedia sistem penyewaan sepeda ini untuk **mengetahui seberapa banyak sepeda yang harus di distribusikan agar sepeda mereka dapat digunakan secara efisien sehingga mengurangi kemungkinan adanya kondisi yang mempunyai banyak sepeda namun sedikit yang menggunakannya ataupun sebaliknya.**"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**Goals**\n",
    "\n",
    "Berdasarkan permasalahan yang telah dijelaskan sebelumnya, pihak yang mengadakan program penyewaan sepeda perlu memiliki 'tool' yang dapat digunakan untuk **memprediksi seberapa banyak sepeda yang harus di distribusikan berdasarkan kondisi tertentu**. adanya perbedaan kondisi dan waktu dapat menambah keakuratan dalam memprediksi banyaknya sepeda yang digunakan, yang mana dapat mendistribusikan sepeda secara efisien sehingga penyedia dapat memaksimalkan keuntungan dari penyewaan tersebut."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**Analytics Approach**\n",
    "\n",
    "Jadi, yang perlu dilakukan adalah menganalisis data untuk menemukan pola dari fitur-fitur yang ada, yang membedakan pada kondisi dan waktu tertentu.\n",
    "\n",
    "Selanjutnya, kita akan membangun suatu model untuk memprediksi suatu nilai numerik atau biasa disebut dengan model regresi, model ini dapat digunakan sebagai 'tool' yang membantu penyedia sistem penyewaan sepeda ini untuk memprediksi banyaknya penyewa berdasarkan kondisi dan waktu tertentu."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**Metrics Evaluation**\n",
    "\n",
    "Evaluasi metrik yang kita gunakan yang paling difokuskan adalah RMSE dan RMSLE, namun kita juga akan menggunakan MAE dan MAPE sebagai pembanding pada kasus ini.\n",
    "* RMSE merupakan nilai rataan akar kuadrat dari error, metrik ini dapat digunakan untuk mengukur seberapa jauh nilai prediksi yang dihasilkan berdasarkan rataannya.\n",
    "* RMSLE adalah nilai rataan akar kuadrat dari error yang digunakan khusus jika fitur target ditransformasikan menggunakan perhitungan logaritma, kita dapat mengunakan metrik ini untuk menghitung rasio antara nilai aktual dan prediksinya. nilai dari metrik ini dapat mereduksi error yang ditimbulkan oleh outliers atau perbedaan nilai yang terlalu jauh.\n",
    "* MAE adalah rataan nilai absolut dari error\n",
    "* MAPE adalah rataan persentase error yang dihasilkan oleh model regresi, dengan nilai keakuratan berdasarkan buku dari [Lewis's (1982)](https://www.nrpa.org/globalassets/journals/jlr/2003/volume-35/jlr-volume-35-number-4-pp-441-454.pdf) mengatakan bahwa jika nilai MAPE sebesar 0.10 atau 10% merupakan prediksi yang sangat akurat, 11-20% merupakan prediksi yang baik, 21-50% merupakan prediksi yang masuk akal, dan 51% dan lebih banyak lagi adalah prediksi yang tidak akurat, adapun jika lebih dari 100% merupakan prediksi yang sangat tidak akurat.\n",
    "\n",
    "Semakin kecil nilai RMSLE, MAE, dan MAPE yang dihasilkan, berarti model semakin akurat dalam memprediksi banyaknya penyewa sepeda dengan limitasi fitur yang digunakan. "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### **Importing Libraries**\n",
    "---"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\indon\\miniconda3\\lib\\site-packages\\xgboost\\compat.py:36: FutureWarning: pandas.Int64Index is deprecated and will be removed from pandas in a future version. Use pandas.Index with the appropriate dtype instead.\n",
      "  from pandas import MultiIndex, Int64Index\n"
     ]
    }
   ],
   "source": [
    "from sklearn.model_selection import cross_val_score, KFold, train_test_split, RandomizedSearchCV\n",
    "from sklearn.metrics import mean_squared_error ,mean_squared_log_error, mean_absolute_error, mean_absolute_percentage_error\n",
    "from sklearn.base import BaseEstimator, TransformerMixin\n",
    "from sklearn.pipeline import Pipeline\n",
    "from sklearn.compose import ColumnTransformer, TransformedTargetRegressor\n",
    "import category_encoders as ce\n",
    "from sklearn.preprocessing import OneHotEncoder, StandardScaler\n",
    "from sklearn.linear_model import LinearRegression\n",
    "from sklearn.tree import DecisionTreeRegressor\n",
    "from sklearn.ensemble import RandomForestRegressor\n",
    "from xgboost.sklearn import XGBRegressor\n",
    "from lightgbm import LGBMRegressor\n",
    "import pandas as pd\n",
    "import pickle\n",
    "import numpy as np\n",
    "import seaborn as sns\n",
    "import matplotlib.pyplot as plt\n",
    "%matplotlib inline\n",
    "\n",
    "import warnings\n",
    "warnings.filterwarnings('ignore')\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**Library Version Used**"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "1.0.2 scikit-learn\n",
      "2.4.1 category_encoder\n",
      "1.5.1 xgboost\n",
      "3.3.2 lightgbm\n",
      "1.4.2 pandas\n",
      "1.21.5 numpy\n",
      "0.11.2 seaborn\n",
      "3.5.1 matplotlib\n",
      "4.0 pickle\n",
      "3.9.12 python\n"
     ]
    }
   ],
   "source": [
    "import sklearn\n",
    "import lightgbm\n",
    "import xgboost\n",
    "import matplotlib\n",
    "from platform import python_version\n",
    "\n",
    "lib = {'scikit-learn': sklearn, 'category_encoder': ce, 'xgboost': xgboost,\n",
    "       'lightgbm': lightgbm, 'pandas': pd, 'numpy': np, 'seaborn': sns,\n",
    "       'matplotlib': matplotlib}\n",
    "\n",
    "for i in lib:\n",
    "    print(lib[i].__version__, i)\n",
    "else:\n",
    "    print(pickle.format_version, 'pickle')\n",
    "    print(python_version(), 'python')\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### **Data Understanding**\n",
    "---\n",
    "\n",
    "Sistem penyewaan sepeda sangat berkorelasi dengan kondisi lingkungan dan musim. Misalnya, kondisi cuaca,\n",
    "curah hujan, hari dalam seminggu, musim, jam dalam sehari, dll. hal tersebut dapat mempengaruhi perilaku penyewa apakah seseorang dengan kondisi tersebut ingin menyewa sepeda atau tidak.\n",
    "\n",
    "* Dataset menunjukkan data sewa sepeda per jam selama dua tahun (2011 dan 2012). \n",
    "* Memiliki 11 atribut yang masing-masing menunjukan waktu penyewaan, kondisi lingkungan, dan banyaknya sepeda yang disewa"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "|Atribute|Data Type|Description|\n",
    "|---|---|---|\n",
    "|dteday|Datetime|Date|\n",
    "|hum|Float|Normalized Humadity. The values are divided to 100 (maximum value)|\n",
    "|weathersit|Integer|1: Clear, few clouds, partly cloudy<br>2: Mist, cloudy, mist and broken clouds, mist and few clouds<br>3: Light snow, light rain and thunderstorm and scattered clouds, light rain and scattered clouds<br>4: Heavy rain, ice pallets, thunderstorm and mist, snow and fog|\n",
    "|holiday|Integer|Whether the day is holiday or not (1: holiday, 0: not holiday). References: [2011](https://dchr.dc.gov/page/holiday-schedules-2010-and-2011), [2012](https://dchr.dc.gov/page/holiday-schedules-2012-and-2013)|\n",
    "|season|Integer|Season (1: Winter, 2: Spring, 3: Summer, 4: Fall)|\n",
    "|atemp|Float|Normalized feeling temperature in Celsius. The values are divided to 50 (maximum value)|\n",
    "|temp|Float|Normalized temperature in Celsius. The values are divided to 41 (maximum value)|\n",
    "|hr|Integer|Hour (0 to 23)|\n",
    "|casual|Integer|Count of casual users|\n",
    "|registered|Integer|Count of registered users|\n",
    "|cnt|Integer|Count of hourly total rental bikes including both casual and registered (Target variable)|"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "df=pd.read_csv('data_bike_sharing.csv', parse_dates=['dteday'])\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(12165, 11)"
      ]
     },
     "execution_count": 4,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "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>dteday</th>\n",
       "      <th>hum</th>\n",
       "      <th>weathersit</th>\n",
       "      <th>holiday</th>\n",
       "      <th>season</th>\n",
       "      <th>atemp</th>\n",
       "      <th>temp</th>\n",
       "      <th>hr</th>\n",
       "      <th>casual</th>\n",
       "      <th>registered</th>\n",
       "      <th>cnt</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>2011-12-09</td>\n",
       "      <td>0.62</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>4</td>\n",
       "      <td>0.3485</td>\n",
       "      <td>0.36</td>\n",
       "      <td>16</td>\n",
       "      <td>24</td>\n",
       "      <td>226</td>\n",
       "      <td>250</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>2012-06-17</td>\n",
       "      <td>0.64</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>2</td>\n",
       "      <td>0.5152</td>\n",
       "      <td>0.54</td>\n",
       "      <td>4</td>\n",
       "      <td>2</td>\n",
       "      <td>16</td>\n",
       "      <td>18</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>2011-06-15</td>\n",
       "      <td>0.53</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>2</td>\n",
       "      <td>0.6212</td>\n",
       "      <td>0.62</td>\n",
       "      <td>23</td>\n",
       "      <td>17</td>\n",
       "      <td>90</td>\n",
       "      <td>107</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>2012-03-31</td>\n",
       "      <td>0.87</td>\n",
       "      <td>2</td>\n",
       "      <td>0</td>\n",
       "      <td>2</td>\n",
       "      <td>0.3485</td>\n",
       "      <td>0.36</td>\n",
       "      <td>8</td>\n",
       "      <td>19</td>\n",
       "      <td>126</td>\n",
       "      <td>145</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>2012-07-31</td>\n",
       "      <td>0.55</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>3</td>\n",
       "      <td>0.6970</td>\n",
       "      <td>0.76</td>\n",
       "      <td>18</td>\n",
       "      <td>99</td>\n",
       "      <td>758</td>\n",
       "      <td>857</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "      dteday   hum  weathersit  holiday  season   atemp  temp  hr  casual  \\\n",
       "0 2011-12-09  0.62           1        0       4  0.3485  0.36  16      24   \n",
       "1 2012-06-17  0.64           1        0       2  0.5152  0.54   4       2   \n",
       "2 2011-06-15  0.53           1        0       2  0.6212  0.62  23      17   \n",
       "3 2012-03-31  0.87           2        0       2  0.3485  0.36   8      19   \n",
       "4 2012-07-31  0.55           1        0       3  0.6970  0.76  18      99   \n",
       "\n",
       "   registered  cnt  \n",
       "0         226  250  \n",
       "1          16   18  \n",
       "2          90  107  \n",
       "3         126  145  \n",
       "4         758  857  "
      ]
     },
     "execution_count": 5,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "dteday        datetime64[ns]\n",
       "hum                  float64\n",
       "weathersit             int64\n",
       "holiday                int64\n",
       "season                 int64\n",
       "atemp                float64\n",
       "temp                 float64\n",
       "hr                     int64\n",
       "casual                 int64\n",
       "registered             int64\n",
       "cnt                    int64\n",
       "dtype: object"
      ]
     },
     "execution_count": 6,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df.dtypes"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**Splitting Dataset into Train and Test Set (no needed, for internal test only)**"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [],
   "source": [
    "# df, test = train_test_split(df, random_state=42, test_size=0.02)\n",
    "# print(df.shape)\n",
    "# print(test.shape)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### **EDA & Feature Engineering**\n",
    "---"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**Null and Duplicated Values**"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "dteday        False\n",
       "hum           False\n",
       "weathersit    False\n",
       "holiday       False\n",
       "season        False\n",
       "atemp         False\n",
       "temp          False\n",
       "hr            False\n",
       "casual        False\n",
       "registered    False\n",
       "cnt           False\n",
       "dtype: bool"
      ]
     },
     "execution_count": 8,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df.isna().any()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "False"
      ]
     },
     "execution_count": 9,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df.duplicated().any()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Berdasarkan hasil diatas, tidak ditemukan adanya null values maupun data yang terduplikasi"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**Drop Unused Features**\n",
    "\n",
    "Kita akan menghilangkan fitur yang dirasa tidak perlu digunakan seperti  `casual` dan `registered` yang akan di total dan menjadi fitur target yaitu `cnt`, jika di pertahankan maka dikhawatirkan akan membuat model menjadi bias atau tidak sesuai."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Index(['dteday', 'hum', 'weathersit', 'holiday', 'season', 'atemp', 'temp',\n",
       "       'hr', 'cnt'],\n",
       "      dtype='object')"
      ]
     },
     "execution_count": 10,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df.drop(['casual', 'registered'], axis=1, errors='ignore', inplace=True)\n",
    "df.columns"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**Handling Outlier with Chebyshev’s & Empirical rules**"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 864x504 with 8 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "bp = df.drop('dteday', axis=1).columns.tolist()\n",
    "n = 1\n",
    "\n",
    "plt.figure(figsize=(12,7))\n",
    "for i in bp:\n",
    "    plt.subplot(3,3,n)\n",
    "    df.boxplot(i)\n",
    "    n += 1"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Terlihat bahwa terdapat outlier yang cukup banyak pada fitur `cnt`, namun kita tidak memiliki bukti kuat untuk dapat menghapusnya dengan asumsi data tersebut bukan merupakan suatu kesalahan. penanganan yang akan dilakukan adalah menggunakan pendekatan [Chebyshev’s & Empirical rules](https://www.csus.edu/indiv/s/seria/lecturenotes/chebyshev.htm) menggunakan standar deviasi 3 dari mean, dengan begitu kita akan menyaring dan menggunakan 99.7% data yang tidak menyimpang atau menghilangkan 0.3% data yang menyimpang."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Before (12165, 9)\n",
      "After (11998, 9)\n"
     ]
    }
   ],
   "source": [
    "before = df.shape\n",
    "df = df[np.abs(df['cnt']-df['cnt'].mean())<=(3*df['cnt'].std())]\n",
    "after = df.shape\n",
    "print('Before', before)\n",
    "print('After', after)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**Handling Skewness with Log Transformation**"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 864x576 with 4 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.figure(figsize=(12, 8))\n",
    "\n",
    "plt.subplot(2, 2, 1)\n",
    "plt.boxplot(df['cnt'])\n",
    "plt.title('Before Transformation')\n",
    "\n",
    "plt.subplot(2, 2, 2)\n",
    "plt.boxplot(df['cnt'].apply(lambda x: np.log1p(x)))\n",
    "plt.title('After Transformation')\n",
    "\n",
    "plt.subplot(2, 2, 3)\n",
    "df['cnt'].hist()\n",
    "\n",
    "plt.subplot(2, 2, 4)\n",
    "df['cnt'].apply(lambda x: np.log1p(x)).hist()\n",
    "\n",
    "plt.show()\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Kita akan melakukan penanganan data yang terindikasi sangat miring dan banyak outliers ini menggunakan transformasi logaritma, berdasarkan penelitian dari [sciencedirect.com](https://www.sciencedirect.com/topics/computer-science/log-transformation) metode ini berguna untuk mengompresi sumbu y. Misalnya, jika kita memiliki rentang data yang sangat besar, maka nilai yang lebih kecil dapat kewalahan oleh nilai yang lebih besar. Transformasi log juga kebal outlier dan memungkinkan kita untuk berpotensi mendapatkan distribusi yang ideal.\n",
    "\n",
    "Transformasi Log digunakan karena nilai pada fitur `cnt` tidak dihasilkan karena kesalahan, jadi kita dapat menganggapnya sebagai outlier alami, berdasarkan asumsi tersebut kita tidak dapat menghapus data outlier ini, maka untuk menangani outlier tersebut kita menggunakan transformasi logaritma."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Before Transformation: 1.04506593926659\n",
      "After Transformation: -0.8248840101070347\n"
     ]
    }
   ],
   "source": [
    "original_skew = df['cnt'].skew()\n",
    "df['cnt'] = df['cnt'].apply(lambda x: np.log1p(x))\n",
    "\n",
    "print('Before Transformation:', original_skew)\n",
    "print('After Transformation:', df['cnt'].skew())"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Dikutip dari situs [Oracle](https://docs.oracle.com/cd/E57185_01/CBREG/ch03s02s03s01.html), Nilai kemiringan lebih besar dari 1 atau kurang dari -1 menunjukkan distribusi yang sangat miring. Nilai antara 0,5 dan 1 atau -0,5 dan -1 kemiringan moderat. Nilai antara -0,5 dan 0,5 menunjukkan bahwa distribusinya cukup simetris. \n",
    "\n",
    "Maka bisa disimpulkan berdasarkan hasil nilai kemiringan diatas memiliki -0.83 yang berarti memiliki kemiringan moderat, lebih baik dari sebelum transformasi di angka 1.0377 yang bisa di katakan sangat miring."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**Categorical Features**"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAWAAAAFkCAYAAAAe8OFaAAAAOXRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjUuMSwgaHR0cHM6Ly9tYXRwbG90bGliLm9yZy/YYfK9AAAACXBIWXMAAAsTAAALEwEAmpwYAAARnUlEQVR4nO3df5BdZX3H8c8nm4TfNFC2Bkk0dkopmvJDtxFIKxqFBkRsGaYVC5UOTDoCbeJYMtVpKTK1ncbCMFWqRqXUQlFa0LHMiEYFIgTBDfI7oIBEErJkU4xJJCAh3/5xzpabZc3e3eyz33t236+ZO3vvuWfP8+wF3px99txdR4QAAONvSvYEAGCyIsAAkIQAA0ASAgwASQgwACQhwACQhABjwrB9jO1L69ucQc+dazvq29tTJjiI7Tktc7q0Zfu59dewJG92GA9TsycAjKFjJP1dff82SU9lTWQPnSvpRElrJV2ZOhMURYCBMWZ7n4jYPtx+EfGUJJefEToVSxDYY7bf3/Kt9NH1tpNatv1Gve0PW7Yd1/K5d9neZnu77Xts//Gg47/H9rdsP2P7Rds/t73a9p+37HONpH9r+bRbB8YaYsoH2l5ue7PtDbaX2d7lZMT279v+dr3PC7YftH2hbbfsc2nL1/NO29+0/bykf6yfv8D2fba31HN+wvYNtt9UP/+qJYh6vifWQ7y+5flrRvwPBh2PM2CMhZUt94+XdH/9sXXb45JOqB8/L2m17csk/e2gY/2OpC/Znh0R/1xvmy/pnS37TJf0ZkmfsT0lIj49wvl+TtKv1fd/RdLFkp6U9BlJsn1evU/r2elcSZ+SdKSki4Y45n9JOmjgQf0/kasG7fPr9e0GSQ+PcM6YgDgDxh6LiHV6Zb11ILwnSIohtknS9yTNkvTR+vFVkg5WFbDr622X2R4I2lclzZP0q5KmSXq9pHvr5z5Yz+FcSX/WMq13RIQjYqhv8bdKOkLVmvEL9bYzJcn2/pKuUBXfmyQdKml/SZfX+11g+8ghjvmspKMkHSDpXyT9br39yfoY+6qK91+qWtsdUj3f2+uHawe+hvrrwwTDGTDGynclzZF0fP1t+lsl3VNvO8H2dElvqfddKekkSV314wvrW6t9JB0n6euS1kn6B0kLJL1Gu/57+5ujmOvlEfFDSbL9gKq4z66fO0HSgfX9M+pbK0t6u6Q1g7ZfEhEP1ve32R6I7GGSLlH1XcH9kj4dETtGMWdMQJwBY6wMLEMcLun3JM2QtErV2e5cVeuae7Xs293GMQ+2PUXSzZLOURWzwScNe73qs4b3o5b7A2fAA8dpa15DbHtw0ON/lfQ/qpZLPqhqeeMuSWtt97Q/VUxkBBhj5bst9z9Uf7yrvnVJWlJve0lVlDe17P8HLd9qDywbTImI61QF/eh6v/+QNKN+/sYh5tDu71ZtPQMd/Dmt81ryS+b18SGO+ULrg4h4PiJOV7XWfJKqpYcNkl6r+od0u8HviJ0kCDDGREQ8Jmlj/fD0+uMqVQGWpFPqj731JVorJO2st/297aNsT6+vDPiQXlkHnd4yzHZJv7B9kqRTh5jGT1vuv6n1ioURWKVqjViSLrY93/Zetl9r+3xJj7RzENtn1ldpHCTpTklfVrWUIg1/lj3wdRxie+bIpo8mIcAYSwNnwVMkPR0R6yX1qjrjHIjhSkmKiCcl/VO9ba6q9dEXJf1Y1Q/BXlc/96iqH2RJ0iJVV1B8Q1LfEOP/QK+c3X5K0k7bd4zkC4iIraquipCqJY87VJ3drld1ZcRvtXmouaqWHX5Yz/lZVVd4SNItw3zu9+uP+0naUF+Gdn6b46JBCDDGUusyxCqp+lZcVVwH/P8laxHxUUln1/tuU3WG+7ikayVdUO/zkqT31sfeLukJSR/Qrpe+DRzvaVWRfkK7LjOMSER8VtUZ+7cl/Uyv/I/hJkl/0uZhvqXqcrMfqwrwNlVnz5dI+pthPveTqpZbNg2zHxrO/EkiAMjBGTAAJCHAAJCEAANAEgIMAEk66q3ICxcujFtuGe4KHQBonCGvSe+oM+BNm7jqBsDk0VEBBoDJhAADQBICDABJCDAAJCHAAJCEAANAEgIMAEkIMAAkIcAAkIQAA0ASAgwASQgwACTpqN+GBgCjsXTpUvX19WnmzJlatmxZ9nTaRoABNF5fX5/Wr1+fPY0RYwkCAJIQYABIUnQJwvZTkrZKelnSjojoKTkeADTJeKwBvyMi+FMXADAISxAAkKR0gEPSN22vtr1oqB1sL7Lda7u3v7+/8HQAoHOUDvD8iHizpFMkXWj7bYN3iIjlEdETET3d3d2FpwMAnaNogCPimfrjRklfkTSv5HgA0CTFAmx7P9sHDNyXdLKkh0qNBwBNU/IqiNdI+ortgXH+MyJuKTgeADRKsQBHxJOSji51fABoOi5DA4AkBBgAkhBgAEhCgAEgCQEGgCQEGACSEGAASEKAASAJAQaAJAQYAJIQYABIQoABIMl4/E24Rlu6dKn6+vo0c+ZMLVu2LHs6ACYQAjyMvr4+rV+/PnsaACYgliAAIEljzoDfcvEXU8Y9YNNWdUn6yaataXNY/Yk/TRkXQFmcAQNAEgIMAEkaswSRZef0/Xb5CABjhQAP4+eHn5w9BQATFEsQAJCEM2AUwRtYgOERYBTBG1iA4bEEAQBJCDAAJGEJYoL7yWW/nTLujucOljRVO55bmzKH113y4LiPCYwUZ8AAkIQAA0ASAgwASVgDRhGH7L1T0o76I4ChEGAU8VdHbc6eAtDxWIIAgCQEGACSEGAASEKAASAJAQaAJAQYAJIQYABIQoABIAkBBoAkBBgAkhBgAEjC74IAOgh/zHRyIcBAB+GPmU4uLEEAQBICDABJCDAAJCHAAJCk+A/hbHdJ6pW0PiJOKz0eMBbmf3J+yrjTN0/XFE3R05ufTpnDnX9x57iPOZmNxxnwYklrxmEcAGiUogG2PUvSuyV9vuQ4ANBEpc+Ar5S0VNIv/dO4thfZ7rXd29/fX3g6ANA5igXY9mmSNkbE6t3tFxHLI6InInq6u7tLTQcAOk7JH8LNl3S67VMl7S3pQNvXRsTZBccEkOj2t52YMu72qV2Sre3r1qXM4cSVt4/q84qdAUfERyJiVkTMkfQ+Sd8hvsDuxb6hnfvtVOwb2VPBOOB3QQAd5KX5L2VPAeNoXAIcEbdJum08xgKApuCdcACQhAADQBICDABJCDAAJCHAAJCEAANAEgIMAEkIMAAkIcAAkIQAA0ASAgwASQgwACQhwACQhAADQBICDABJCDAAJCHAAJCEAANAEgIMAEkIMAAkIcAAkIQAA0ASAgwASQgwACQhwACQhAADQBICDABJCDAAJCHAAJCEAANAEgIMAEkIMAAkIcAAkIQAA0ASAgwASQgwACQhwACQhAADQBICDABJCDAAJCHAAJCEAANAEgIMAEmmZk8AAPbUjIhdPjYFAQbQeGe/vDN7CqPCEgQAJCHAAJCEAANAEgIMAEmKBdj23rbvsX2/7Ydtf6zUWADQRCWvgnhR0oKI2GZ7mqQ7bH89Ir5XcEwAaIxiAY6IkLStfjitvjXrIj0AKKjoGrDtLtv3SdooaUVE3D3EPots99ru7e/vLzkdAOgoRQMcES9HxDGSZkmaZ3vuEPssj4ieiOjp7u4uOR0A6CjjchVERGyWdJukheMxHgA0QcmrILptz6jv7yPpXZIeLTUeADRNyasgDpX077a7VIX+hoi4ueB4ANAoJa+CeEDSsaWODwBNxzvhACAJAQaAJAQYAJIQYABIQoABIAkBBoAkBBgAkrQVYNuL29kGAGhfu2fAHxhi27ljOA8AmHR2+04422dJer+kN9j+WstTB0j635ITA4CJbri3Iq+StEHSIZIub9m+VdIDpSYFAJPBbgMcEWslrZV0/PhMBwAmj3Z/CHeG7R/Z/pntLba32t5SenIAMJG1+9vQlkl6T0SsKTkZAJhM2r0K4lniCwBjq90z4F7bX5b0VVV/bl6SFBE3lZgUAEwG7Qb4QEnPSzq5ZVtIIsAAMErtBniKpMX1H9eU7YO062VpAIARancN+KiB+EpSRPxU/LkhANgj7QZ4Sn3WK0myfbDK/kFPAJjw2o3o5ZJW2f5vVWu/fyTp48VmBQCTQFsBjogv2u6VtECSJZ0REY8UnRkATHBtLyPUwSW6ADBG+IXsAJCEAANAEgIMAEkIMAAkIcAAkIQAA0ASAgwASQgwACQhwACQhAADQBICDABJCDAAJCHAAJCEAANAEgIMAEkIMAAkIcAAkIQAA0ASAgwASQgwACQhwACQhAADQBICDABJCDAAJCHAAJCkWIBtz7Z9q+01th+2vbjUWADQRFMLHnuHpA9HxL22D5C02vaKiHik4JgA0BjFzoAjYkNE3Fvf3yppjaTDSo0HAE0zLmvAtudIOlbS3UM8t8h2r+3e/v7+8ZgOAHSE4gG2vb+kGyUtiYgtg5+PiOUR0RMRPd3d3aWnAwAdo2iAbU9TFd/rIuKmkmMBQNOUvArCkr4gaU1EXFFqHABoqpJnwPMlnSNpge376tupBccDgEYpdhlaRNwhyaWODwBNxzvhACAJAQaAJAQYAJIQYABIQoABIAkBBoAkBBgAkhBgAEhCgAEgCQEGgCQEGACSEGAASEKAASAJAQaAJAQYAJIQYABIQoABIAkBBoAkBBgAkhBgAEhCgAEgCQEGgCQEGACSEGAASEKAASAJAQaAJAQYAJIQYABIQoABIAkBBoAkBBgAkhBgAEhCgAEgCQEGgCQEGACSEGAASEKAASAJAQaAJAQYAJIQYABIQoABIAkBBoAkBBgAkhBgAEhCgAEgCQEGgCQEGACSEGAASFIswLavtr3R9kOlxgCAJit5BnyNpIUFjw8AjVYswBGxUtJzpY4PAE2XvgZse5HtXtu9/f392dMBgHGTHuCIWB4RPRHR093dnT0dABg36QEGgMmKAANAkpKXoV0v6S5JR9heZ/u8UmMBQBNNLXXgiDir1LEBYCJgCQIAkhBgAEhCgAEgCQEGgCQEGACSEGAASEKAASAJAQaAJAQYAJIQYABIQoABIAkBBoAkBBgAkhBgAEhCgAEgCQEGgCQEGACSEGAASEKAASAJAQaAJAQYAJIQYABIQoABIAkBBoAkBBgAkhBgAEhCgAEgCQEGgCQEGACSEGAASEKAASAJAQaAJAQYAJIQYABIQoABIAkBBoAkBBgAkhBgAEhCgAEgCQEGgCQEGACSEGAASEKAASAJAQaAJAQYAJIQYABIQoABIAkBBoAkBBgAkhQNsO2Fth+z/bjtvy45FgA0TbEA2+6SdJWkUyS9UdJZtt9YajwAaJqSZ8DzJD0eEU9GxC8kfUnSewuOBwCN4ogoc2D7TEkLI+L8+vE5kt4aERcN2m+RpEX1wyMkPVZkQnvmEEmbsifRQLxuo8PrNjqd/LptioiFgzdOLTigh9j2qtpHxHJJywvOY4/Z7o2Inux5NA2v2+jwuo1OE1+3kksQ6yTNbnk8S9IzBccDgEYpGeDvSzrc9htsT5f0PklfKzgeADRKsSWIiNhh+yJJ35DUJenqiHi41HiFdfQSSQfjdRsdXrfRadzrVuyHcACA3eOdcACQhAADQBICvBu2r7a90fZD2XNpEtuzbd9qe43th20vzp5TE9je2/Y9tu+vX7ePZc+pKWx32f6B7Zuz5zISBHj3rpH0qounMawdkj4cEUdKOk7ShbwNvS0vSloQEUdLOkbSQtvH5U6pMRZLWpM9iZEiwLsRESslPZc9j6aJiA0RcW99f6uq/zAOy51V54vKtvrhtPrGT8mHYXuWpHdL+nz2XEaKAKMo23MkHSvp7uSpNEL9rfR9kjZKWhERvG7Du1LSUkk7k+cxYgQYxdjeX9KNkpZExJbs+TRBRLwcEceoeufoPNtzk6fU0WyfJmljRKzOnstoEGAUYXuaqvheFxE3Zc+naSJis6TbxM8ghjNf0um2n1L1GxcX2L42d0rtI8AYc7Yt6QuS1kTEFdnzaQrb3bZn1Pf3kfQuSY+mTqrDRcRHImJWRMxR9esOvhMRZydPq20EeDdsXy/pLklH2F5n+7zsOTXEfEnnqDobua++nZo9qQY4VNKtth9Q9btUVkREoy6rwsjwVmQASMIZMAAkIcAAkIQAA0ASAgwASQgwACQhwACQhAADQJL/A/w+5avrufOZAAAAAElFTkSuQmCC\n",
      "text/plain": [
       "<Figure size 360x360 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 216x360 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 360x360 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAABFQAAAFkCAYAAAAOv29aAAAAOXRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjUuMSwgaHR0cHM6Ly9tYXRwbG90bGliLm9yZy/YYfK9AAAACXBIWXMAAAsTAAALEwEAmpwYAAAfX0lEQVR4nO3dfZBlZ10n8O8vCYG8TEgwExpNYtDFrJalgAOKaFgTC8JbeDEorC+AWHGpRYOLtCBbrhZllTbiqrUWmuVFVhBxA2EBJcCqkbUUJOE1kKCIQdKTm0wUQiZAQsKzf9ybcWYy/XJy5/S5p+fzqeo6t2/f+5xv3eqZ595vP+ecaq0FAAAAgM07augAAAAAAGOjUAEAAADoSKECAAAA0JFCBQAAAKAjhQoAAABARwoVAAAAgI4UKgDA6FTVFVXVquq6obMAAEcmhQoAAABARwoVAOCIUlXHDZ0BABg/hQoAMGpV9d1V9TdV9aWq+nhVPXZ2/1mzw4JaVb28qn6tqiZJdg8cGQDYBqq1NnQGAIBOquqKJI9JcluSSnL8fj++Lck3JtmR5J9m930+ySmz27e01k7ekqAAwLZlhQoAMGYnJHl9kgckefl+9z3+oMfdP8lzMi1ZHr1V4QCA7UuhAgCM2Z1JfqG19vkkb9rv/jMOetzlrbXXt9b2ttY+sXXxAIDtSqECAIzZja21W2e3v7Lf/fc96HEf36I8AMARQqECAIzZnfvdXu/EcF9Z52cAAJ0pVAAAAAA6UqgAAAAAdKRQAQAAAOioWlvvcGMAAAAADmaFCgAAAEBHChUAAACAjhQqAAAAAB0pVAAAAAA6OmboAPs7//zz2+WXXz50DAAAAIC71aHuXKgVKjfffPPQEQAAAAA2tFCFCgAAAMAYKFQAAAAAOlKoAAAAAHTUa6FSVSdX1aVVdW1VXVNVj+pzfwAAAABboe+r/Px2kstbaxdW1bFJju95fwAAAAC9661QqaqTkpyT5DlJ0lq7I8kdfe0PAAAAYKv0ecjPNyXZk+R1VfXhqnp1VZ1w8IOq6qKqurKqrtyzZ0+PcQAAAAAOjz4LlWOSPDzJq1prD0tyW5KXHPyg1tolrbVdrbVdO3fu7DEOAAAAwOHRZ6FyfZLrW2sfmH1/aaYFCwAAAMCo9VaotNYmST5XVWfP7jovySf72h8AAADAVun7Kj8/k+SNsyv8fCbJc3veHwAAAEDvei1UWmsfSbKrz30AAAAAbLW+V6gAAADca8vLy5lMJllaWsrKysrQcQD2UagAAAALazKZZHV1degYAPfQ51V+AAAAALYlhQoAAABARw75AQAAevHqt9409xhf3HvXvu084/3U00+bO8tmOOcLHDkUKgAAAIeJc77AkcMhPwAAAAAdKVQAAAAAOnLIDwAAsLBOOGnnAds+vevNN889xpf2fm3fdt7xHv8jp86dB+iPQgUAAFhYP3DBLw4dAeCQHPIDAAAA0JEVKgAAcIRwSV+Aw0ehAkDvvIEHWAwu6du/k07cecB2DMzTcO8oVADonTfwABwpnvHElw0doTPzNNw7zqECAAAA0JEVKgAAMAI/e9nn5h5jz947923nHe93nnbG3HkAxkyhAgAAMFIffN1Nc49x+xfv2redZ7xHPPe0ubPAmChUANjQcy87f67n37j3q7Pt6txjve5pl8/1fAAAOBycQwUAAACgIytUAADgCHGfk049YAvAvadQAQCAI8RZT3nx0BEAtg2FCgDAJiwvL2cymWRpaSkrKytDxwE4bE45cecBW2BzFCoA9O6YkypJm21hnCaTSVZXV4eOAXDYPe+8Xxw6AoySQgWA3p32FNMNDMGqGgDoj3e4AADblFU1/VNaARy5FCoAwBHhSZe+ca7nf2XvrUmS3XtvnXusd174o3M9f7saYzmhtAI4cilUAABYCMoJAMZEoQIAsICefOnb5h7jy3tvS5Ls3nvb3OO948Knzp1n0Vz4lg/NPcYte29Pktyw9/a5xrv0hx4+dxYAtpZCBQAOYYyHHgDAkcAczaJQqADAITj0gIPVjhMP2HKgp17653OPsXfvl5Mku/d+ee7x3nbheXPnARaTOZpFoVABANiE+z75/KEjAAALRKECwLbzhLe9aO4x7rjt5iTJ7ttunnu8P3vqK+fOAyymo3accsAWgCOHQgUAYJuqHTsO2C66o3bcP1+bbcdixwUXDR0BRuW635rMPcadX7hr33be8c564dLceThyKVQAgEE4qWD/7vfkpw4doZPjn/ysoSMAwKYpVACAQTipIAAwZgoVADiUHcemZlsAADiYQgUADuHYp37z0BEW2hPfcsncY9y+95Ykye69t8w93p/+kPNYABwpvu64Uw/YwlAUKgAAAIzGix710qEjQJLkqKEDAAAAAIyNQgUAAACgI4f8AACDqB0nHLAFABiTXguVqrouya1J7kpyZ2ttV5/7A4Aj1fLyciaTSZaWlrKysjJ0nE059oL/MHQEAOjdGOdoNmcrVqj8QGvt5i3YDwAcsSaTSVZXV4eOAQAcxBy9fTmHCgAAAEBHfRcqLcl7quqqqrroUA+oqouq6sqqunLPnj09xwEAAACYX9+H/Dy6tba7qk5L8t6qura19r79H9BauyTJJUmya9eu1nMegG3BsbjbzxMve8Vcz7997+eTJLv3fn7usf70aS+e6/kAsJ1MXnntXM+/6/Nf3bedd6ylF/37uZ7P4dVrodJa2z3b3lRVlyV5ZJL3rf8sADbiWFwAABhWb4f8VNUJVbXj7ttJHpvk6r72BwAAALBV+lyh8sAkl1XV3fv5o9ba5T3uDwAAAGBL9FaotNY+k+Q7+xofAPg3ddJxB2wBgMVw6vEPOGDL9tH3SWkBgC1w7FMeMXQEAOAQXvrInxk6Aj1RqABssV998+PmHuNf9945267OPd7LfuTdc+cBAIAjTW8npQUAAADYrhQqAAAAAB0pVAAAAAA6cg4VAAAAYJ/l5eVMJpMsLS1lZWVl6DgLS6ECAAAA7DOZTLK6ujp0jIWnUAEYofudWEnabAsAAGw1hQrACD3siUcPHQEAAI5oChUAAADYJm787b+de4y7vvCVfdt5x3vgxY+aO8+icpUfAAAAgI4UKgAAAAAdKVQAAAAAOnIOFQAAAGCfnceffMCWQ1OoAAAAAPu89FHPHTrCKDjkBwAAAKAjhQoAAABARwoVAAAAgI4UKgAAAAAdKVQAAAAAOlKoAAAAAHSkUAEAAADoSKECAAAA0JFCBQAAAKAjhQoAAABAR8cMHQAAAADg3lpeXs5kMsnS0lJWVla2bL8KFQAAAGC0JpNJVldXt3y/DvkBAAAA6EihAgAAANCRQ34AAACAwdz0P9411/Pv+sKX9m3nHeu0Fzx+04+1QgUAAACgI4UKAAAAQEcKFQAAAICOnEMFAAAAGK2dJ5x0wHarKFQAAACA0frFRz9jkP065AcAAACgI4UKAAAAQEcKFQAAAICOFCoAAAAAHSlUAAAAADpSqAAAAAB01HuhUlVHV9WHq+qdfe8LAAAAYCscswX7uDjJNUlO2oJ9Adwry8vLmUwmWVpaysrKytBxAACABdfrCpWqOj3JE5O8us/9AMxrMplkdXU1k8lk6CgAAMAI9H3Iz28lWU7ytbUeUFUXVdWVVXXlnj17eo4DAAAAML/eCpWqelKSm1prV633uNbaJa21Xa21XTt37uwrDgAAAMBh0+cKlUcnuaCqrkvyx0nOrao39Lg/AAAAgC3RW6HSWntpa+301tpZSZ6Z5C9aaz/W1/4AAAAAtspWXOUHoFe//4ePm3uMW269c7ZdnWu8n/7xd8+dBQAAWHxbUqi01q5IcsVW7AsAAACgb31f5QcAAABg21GoAAAAAHSkUAEAAADoSKECAAAA0JGr/AAkOeHEStJmWwAAgPVt+0JleXk5k8kkS0tLWVlZGToOsKAe89ijh44AAACMyLYvVCaTSVZXV4eOAQAAAGwjzqECAAAA0JFCBQAAAKCjhT7kZ8+r3jD3GHfdcuu+7bzj7Xz+j82dBwAAABg/K1QAAAAAOlKoAAAAAHSkUAEAAADoaKHPoXI47Dz+xAO2AAAAAPPa9oXKy8553NARAAAAgG3GIT8AAAAAHSlUAAAAADpSqAAAAAB0pFABAAAA6EihAgAAANCRQgUAAACgI4UKAAAAQEcKFQAAAICOFCoAAAAAHSlUAAAAADpSqAAAAAB0pFABAAAA6EihAgAAANCRQgUAAACgI4UKAAAAQEcKFQAAAICOFCoAAAAAHSlUAAAAADpSqAAAAAB0pFABAAAA6GhThUpVXbyZ+5jf8vJyfuInfiLLy8tDRwEAAADWsNkVKs8+xH3POYw5mJlMJlldXc1kMhk6CgAAALCGY9b7YVU9K8l/TPLgqnr7fj/akeRf+gwGAAAAsKjWLVSS/E2SG5KcmuSV+91/a5KP9RUKAAAAYJGtW6i01j6b5LNJHrU1cQAAAAAW30YrVJIkVfX0JL+e5LQkNftqrbWTesw2Oje+6hVzj3HXLZ/ft513vAc+/8Vz5wEAAADuaVOFSpKVJE9urV2z2YGr6n5J3pfkvrP9XNpa+2/dIwIAAAAsls0WKjd2KVNmbk9ybmttb1XdJ8lfV9W7Wmvv7zgOAAAAwELZbKFyZVW9OcnbMi1KkiSttbeu9YTWWkuyd/btfWZf7d7FBAAAAFgcmy1UTkrypSSP3e++lmTNQiVJquroJFcl+XdJfre19oFDPOaiJBclyZlnnrnJOAAAAADD2WyhclSSi1trX0iSqjolB15G+ZBaa3cleWhVnZzksqr69tba1Qc95pIklyTJrl27jvgVLDuPP+6ALQAAALB4NluofMfdZUqStNY+X1UP2+xOWmtfqKorkpyf5OoNHn5Ee+k5jxg6AgAAALCBozb7uNmqlCRJVT0gG5QxVbVztjIlVXVckh9Mcu29zAkAAACwMDa7QuWVSf6mqi7N9NwpP5zkVzd4zoOSvH52HpWjkvxJa+2d9zopAAAAwILYVKHSWvtfVXVlknOTVJKnt9Y+ucFzPpZk04cFAdvH8vJyJpNJlpaWsrKyMnQcAACAw26zK1QyK1DWLVEAkmQymWR1dXXoGAAAAL3Z7DlUAAAAAJhRqAAAAAB0pFABAAAA6GjT51ABjgxvft35c4+x94tfnW1X5x7vR557+dx5AAAADjcrVAAAAAA6UqgAAAAAdOSQH+a2vLycyWSSpaWlrKysDB0HAAAAeqdQYW6TySSrq6tDx2CBnHhiJWmzLQAAwPajUAEOuyee578WAABge3MOFQAAAICOFCoAAAAAHSlUAAAAADpyooMj3D//zoVzj3HnF26ZbW+Ye7wzf/bSufMAAABA36xQAQAAAOhIoQIAAADQkUIFAAAAoCOFCgAAAEBHTkrL3E49/qgDtgAAALDdKVSY289/746hIwAAAMCWsqQAAAAAoCOFCgAAAEBHChUAAACAjhQqAAAAAB0pVAAAAAA6UqgAAAAAdKRQAQAAAOhIoQIAAADQkUIFAAAAoCOFCgAAAEBHChUAAACAjhQqAAAAAB0pVAAAAAA6OmboALDVlpeXM5lMsrS0lJWVlaHjAAAAMEIKFY44k8kkq6urQ8cAAABgxBzyAwAAANCRQgUAAACgI4UKAAAAQEfOocKofPD3nzz3GLff8uXZdvfc4z3ip98xdx4AAADGxwoVAAAAgI4UKgAAAAAd9VaoVNUZVfWXVXVNVX2iqi7ua18AAAAAW6nPc6jcmeRFrbUPVdWOJFdV1Xtba5/scZ8AAAAAveutUGmt3ZDkhtntW6vqmiTfkEShwqBOOaEO2AIAAEBXW3KVn6o6K8nDknzgED+7KMlFSXLmmWduRRyOcM97zP2GjgAAAMDI9X5S2qo6MclbkrywtfbFg3/eWruktbartbZr586dfccBAAAAmFuvhUpV3SfTMuWNrbW39rkvAAAAgK3S51V+KslrklzTWvvNvvYDAAAAsNX6XKHy6CQ/nuTcqvrI7OsJPe4PAAAAYEv0eZWfv07iMipwGCwvL2cymWRpaSkrKytDxwEAADjibclVfoD5TCaTrK6uDh0DAACAmd6v8gMAAACw3VihAj1792vmP3XQl754x2y7e67xHve8P5s7CwAAAFaoAAAAAHSmUAEAAADoSKECAAAA0JFzqMAI3P+EJKnZFgAAgKEpVGAEfvjcY4eOAAAAwH4c8gMAAADQkUIFAAAAoCOFCgAAAEBHChUAAACAjhQqAAAAAB0pVAAAAAA6UqgAAAAAdKRQAQAAAOhIoQIAAADQkUIFAAAAoCOFCgAAAEBHChUAAACAjhQqAAAAAB0pVAAAAAA6UqgAAAAAdKRQAQAAAOhIoQIAAADQkUIFAAAAoCOFCgAAAEBHChUAAACAjhQqAAAAAB0pVAAAAAA6UqgAAAAAdKRQAQAAAOhIoQIAAADQkUIFAAAAoCOFCgAAAEBHChUAAACAjhQqAAAAAB0pVAAAAAA6UqgAAAAAdKRQAQAAAOhIoQIAAADQUW+FSlW9tqpuqqqr+9oHAAAAwBD6XKHyB0nO73F8AAAAgEH0Vqi01t6X5F/7Gh8AAABgKIOfQ6WqLqqqK6vqyj179gwdBwAAAGBDgxcqrbVLWmu7Wmu7du7cOXQcAAAAgA0NXqgAAAAAjI1CBQAAAKCjPi+b/KYkf5vk7Kq6vqqe19e+AAAAALbSMX0N3Fp7Vl9jAwAAAAzJIT8AAAAAHSlUAAAAADpSqAAAAAB0pFABAAAA6EihAgAAANCRQgUAAACgI4UKAAAAQEcKFQAAAICOFCoAAAAAHSlUAAAAADpSqAAAAAB0pFABAAAA6EihAgAAANCRQgUAAACgI4UKAAAAQEcKFQAAAICOFCoAAAAAHSlUAAAAADpSqAAAAAB0pFABAAAA6EihAgAAANCRQgUAAACgI4UKAAAAQEcKFQAAAICOFCoAAAAAHSlUAAAAADpSqAAAAAB0pFABAAAA6EihAgAAANCRQgUAAACgI4UKAAAAQEcKFQAAAICOFCoAAAAAHSlUAAAAADpSqAAAAAB0pFABAAAA6EihAgAAANCRQgUAAACgI4UKAAAAQEcKFQAAAICOFCoAAAAAHfVaqFTV+VX1qar6dFW9pM99AQAAAGyV3gqVqjo6ye8meXySb0vyrKr6tr72BwAAALBV+lyh8sgkn26tfaa1dkeSP07ylB73BwAAALAlqrXWz8BVFyY5v7X2U7PvfzzJd7fWXnDQ4y5KctHs27OTfKqHOKcmubmHcfsytrzJ+DKPLW8yvsxjy5uML7O8/Rtb5rHlTcaXeWx5k/FlHlveZHyZx5Y3GV/mseVNxpdZ3v6NLXOfeW9urZ1/8J3H9LSzJKlD3HeP9qa1dkmSS3rMkaq6srW2q899HE5jy5uML/PY8ibjyzy2vMn4Msvbv7FlHlveZHyZx5Y3GV/mseVNxpd5bHmT8WUeW95kfJnl7d/YMg+Rt89Dfq5PcsZ+35+eZHeP+wMAAADYEn0WKh9M8pCqenBVHZvkmUne3uP+AAAAALZEb4f8tNburKoXJHl3kqOTvLa19om+9reBXg8p6sHY8ibjyzy2vMn4Mo8tbzK+zPL2b2yZx5Y3GV/mseVNxpd5bHmT8WUeW95kfJnHljcZX2Z5+ze2zFuet7eT0gIAAABsV30e8gMAAACwLSlUAAAAADra1oVKVZ1fVZ+qqk9X1UuGzrORqnptVd1UVVcPnWUzquqMqvrLqrqmqj5RVRcPnWkjVXW/qvq7qvroLPOvDJ1pM6rq6Kr6cFW9c+gsm1FV11XVx6vqI1V15dB5NlJVJ1fVpVV17ez3+VFDZ1pPVZ09e23v/vpiVb1w6Fzrqaqfm/2bu7qq3lRV9xs603qq6uJZ1k8s6mt7qDmjqh5QVe+tqn+YbU8ZMuPB1sj8jNnr/LWqWqhLM66R9xWz/ys+VlWXVdXJA0a8hzUyv3yW9yNV9Z6q+vohM+5vvfc+VfXzVdWq6tQhsh3KGq/vL1fV6n7/Jz9hyIwHW+s1rqqfmb1P/kRVrQyV72BrvMZv3u/1va6qPjJgxHtYI/NDq+r9d78XqqpHDplxf2vk/c6q+tvZ+7d3VNVJQ2bc31qfORZ5zlsn80LOeevkXdg5b53MWzvntda25VemJ8L9xyTflOTYJB9N8m1D59og8zlJHp7k6qGzbDLvg5I8fHZ7R5K/H8FrXElOnN2+T5IPJPmeoXNtIvd/SfJHSd45dJZN5r0uyalD5+iQ9/VJfmp2+9gkJw+dqUP2o5NMknzj0FnWyfgNSf4pyXGz7/8kyXOGzrVO3m9PcnWS4zM9efv/TfKQoXMdIuc95owkK0leMrv9kiS/PnTOTWT+1iRnJ7kiya6hM24i72OTHDO7/esjeY1P2u/2zyb5vaFzrpd3dv8ZmV7Y4LOLNJ+s8fr+cpKfHzpbx8w/MPu/7b6z708bOudGvxP7/fyVSX5p6JybeI3fk+Txs9tPSHLF0Dk3yPvBJI+Z3f7JJC8fOud+2Q75mWOR57x1Mi/knLdO3oWd89bJvKVz3nZeofLIJJ9urX2mtXZHkj9O8pSBM62rtfa+JP86dI7Naq3d0Fr70Oz2rUmuyfSD08JqU3tn395n9rXQZ2auqtOTPDHJq4fOsh3N/gJzTpLXJElr7Y7W2hcGDdXNeUn+sbX22aGDbOCYJMdV1TGZFhW7B86znm9N8v7W2pdaa3cm+askTxs40z2sMWc8JdOCMLPtU7cy00YOlbm1dk1r7VMDRVrXGnnfM/u9SJL3Jzl9y4OtY43MX9zv2xOyQPPeOu99/nuS5SxQ1mR879WSNTM/P8mvtdZunz3mpi0Ptob1XuOqqiQ/nORNWxpqA2tkbknuXuVx/yzQvLdG3rOTvG92+71JfmhLQ61jnc8cCzvnrZV5Uee8dfIu7Jy3TuYtnfO2c6HyDUk+t9/312fBP+yPWVWdleRhma74WGg1PXzmI0luSvLe1tqiZ/6tTN9Ufm3gHF20JO+pqquq6qKhw2zgm5LsSfK6mh5W9eqqOmHoUB08Mwv2xvJgrbXVJL+R5J+T3JDkltbae4ZNta6rk5xTVV9XVcdn+pfFMwbOtFkPbK3dkEzfaCQ5beA8291PJnnX0CE2o6p+tao+l+RHk/zS0HnWU1UXJFltrX106CwdvGC2xPy1i3TYwTq+Jcn3V9UHquqvquoRQwfapO9PcmNr7R+GDrIJL0zyitm/u99I8tJh42zo6iQXzG4/Iws67x30mWMUc96YPicl6+Zd2Dnv4MxbOedt50KlDnHfQv2VY7uoqhOTvCXJCw9qBBdSa+2u1tpDM21YH1lV3z5wpDVV1ZOS3NRau2roLB09urX28CSPT/Kfq+qcoQOt45hMl72+qrX2sCS3ZbpsdOFV1bGZvvn530NnWc/sw8VTkjw4ydcnOaGqfmzYVGtrrV2T6bLW9ya5PNNDRu9c90kccarqZZn+Xrxx6Cyb0Vp7WWvtjEzzvmDoPGuZlZgvy4KXPgd5VZJvTvLQTEvjVw6aZnOOSXJKku9J8uIkfzJb/bHonpUF/yPCfp6f5Odm/+5+LrOVsAvsJzN9z3ZVpodP3DFwnnsY22eOZHyZ18q7yHPeoTJv5Zy3nQuV63Ngs3p6Fmip3XZRVffJ9Bf4ja21tw6dp4vZYR1XJDl/2CTrenSSC6rqukwPWzu3qt4wbKSNtdZ2z7Y3Jbks00PwFtX1Sa7fb6XSpZkWLGPw+CQfaq3dOHSQDfxgkn9qre1prX01yVuTfO/AmdbVWntNa+3hrbVzMl0WPYa/hibJjVX1oCSZbRdmGf92UlXPTvKkJD/aZgdpj8gfZYGW8h/CN2davn50NvednuRDVbU0aKp1tNZunP2x5mtJ/mcWe8672/VJ3jo7FPrvMl0FuzAn/z2U2SGjT0/y5qGzbNKzM53vkukfPhb696K1dm1r7bGtte/KtLT6x6Ez7W+NzxwLPeeN7XPSWnkXec7bxGvc+5y3nQuVDyZ5SFU9ePZX3GcmefvAmbaV2V8yXpPkmtbabw6dZzOqaufdZ6euquMy/aB37aCh1tFae2lr7fTW2lmZ/g7/RWttYf+ynyRVdUJV7bj7dqYns1rYK1e11iZJPldVZ8/uOi/JJweM1MVY/lL3z0m+p6qOn/2/cV6mx7kurKo6bbY9M9M38GN4nZPpPPfs2e1nJ/k/A2bZlqrq/CS/kOSC1tqXhs6zGVX1kP2+vSCLPe99vLV2WmvtrNncd32mJx2cDBxtTXd/oJt5WhZ4ztvP25KcmyRV9S2ZnpD95iEDbcIPJrm2tXb90EE2aXeSx8xun5sFL+b3m/eOSvJfk/zesIn+zTqfORZ2zhvb56S18i7ynLdO5q2d8w4+S+12+sr0uPe/z7RhfdnQeTaR902ZLhX9aqZvIJ43dKYN8n5fpodRfSzJR2ZfTxg61waZvyPJh2eZr86CnSV+g+z/ISO4yk+m5yT56OzrEyP5t/fQJFfOfi/eluSUoTNtIvPxSf4lyf2HzrLJvL+S6YR2dZI/zOzKEov6leT/ZVqsfTTJeUPnWSPjPeaMJF+X5M8zfeP+50keMHTOTWR+2uz27UluTPLuoXNukPfTmZ6j7e55b2GumLNO5rfM/u19LMk7Mj1p3+BZ18p70M+vy2Jd5edQr+8fJvn47PV9e5IHDZ1zE5mPTfKG2e/Fh5KcO3TOjX4nkvxBkv80dL4Or/H3JblqNo98IMl3DZ1zg7wXZ/q56e+T/FqSGjrnfnkP+Zljkee8dTIv5Jy3Tt6FnfPWybylc17NwgAAAACwSdv5kB8AAACAXihUAAAAADpSqAAAAAB0pFABAAAA6EihAgAAANCRQgUAAACgI4UKAAAAQEf/HxisQ+BzYOGCAAAAAElFTkSuQmCC\n",
      "text/plain": [
       "<Figure size 1116x360 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "time_attr = ['weathersit', 'holiday', 'season', 'hr']\n",
    "\n",
    "for i in time_attr:\n",
    "    l = len(df[i].unique())\n",
    "    if l == 2:\n",
    "        aspect = .6\n",
    "    elif l == 4:\n",
    "        aspect = 1\n",
    "    elif l == 7:\n",
    "        aspect = 1.6\n",
    "    elif l == 12:\n",
    "        aspect = 2.1  \n",
    "    elif l == 24:\n",
    "        aspect = 3.1\n",
    "    elif l == 31:\n",
    "        aspect = 4.1\n",
    "        \n",
    "    sns.catplot(x=i, y='cnt', data=df,\n",
    "                kind='bar', height=5, aspect=aspect)\n",
    "    plt.title(i, size=15, weight='bold')\n",
    "    plt.xlabel(None)\n",
    "    plt.show()\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**Numerical Features**"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAsgAAAFkCAYAAAA9nc1+AAAAOXRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjUuMSwgaHR0cHM6Ly9tYXRwbG90bGliLm9yZy/YYfK9AAAACXBIWXMAAAsTAAALEwEAmpwYAAA9tklEQVR4nO3dd5zdVYH+8efcMnd6r5makGTSE5KBUJUaiiiCoKCCqIjusmtBd3V1V91i2/XHKmtBbIi6oCiyIKChCQECSQjpvc5Mpvc+t53fHzPEm5Aymcyd7713Pu/X675y+zyTb2bmyZnzPcdYawUAAABghMvpAAAAAEAsoSADAAAAESjIAAAAQAQKMgAAABCBggwAAABEoCADAAAAESjIABBjjDF/McZYY8wBp7MAwFREQQYAAAAiUJABAACACBRkAIhhxpjlxphXjDEDxpjNxpgVo/dfNDoNwxpjbot4/v1v3h9x31cjnnuJMea50fd7Y/T984wxvzXG9BljdhtjbnHgUwWAmOFxOgAA4LjyJT0nKXX09gJJjxhjKk/jPR+WlDt6fYmkxyTtl7R89L6Zkn5hjHndWrvtND4OAMQtRpABIHalSfqFRgrtv0fcd9VpvOdaSXmS/mv0dqGkaRopxjeM3mckXX8aHwMA4hoFGQBiV1DS5621nZIejLi//DTe8z+ttR0aGZl+0/3W2r2SHp+gjwEAcY2CDACxq9la2zt6fSjift8JXuM+yXvWHuP9aiXJWusf48cAgIRGQQaA2BWMuG6Pemw44npyxPWqU3jPE90HAFMWBRkA4lN9xPUVZsQlks53KhAAJAoKMgDEIWttnaQ1ozevk9Qj6VkdOXUCADAOFGQAiF83a6QU90tqk/QZSb91NBEAJABj7dHT2gAAAICpixFkAAAAIAIFGQAAAIhAQQYAAAAiUJABAACACB6nA5yqK6+80v7pT39yOgYAAADinznWnXE3gtzW1uZ0BAAAACSwuCvIAAAAQDRRkAEAAIAIFGQAAAAgAgUZAAAAiEBBBgAAACJQkAEAAIAIFGQAAAAgAgUZAAAAiEBBBgAAACJQkAEAAIAIFGQAAAAgAgUZAAAAiEBBBgAAACJQkAEAp6S8olLGmHFdyisqnY4PACflcToAACC+1NfV6u6VO8f12rtWVE9wGgCYeIwgAwAAABEoyAAAAEAECjIAAAAQgYIMAAAARKAgAwAAABEoyAAAAEAECjIAAAAQgYIMAAAARKAgAwAAABEoyAAAAEAECjIAAAAQgYIMAAAARKAgAwAAABEoyAAAAEAECjIAAAAQgYIMAAAARKAgAwAAABGiVpCNMeXGmOeNMduNMVuNMZ86xnOMMeYeY8weY8wmY8zSaOUBAABIZOUVlTLGjPtSXlHp9KcQMzxRfO+gpM9aa9cbYzIkvW6Medpauy3iOVdJmjV6WS7ph6N/AgCAOFdeUan6utpxvbasvEJ1tQcnOFFiq6+r1d0rd4779XetqJ7ANPEtagXZWtsoqXH0eq8xZrukUkmRBflaSQ9Ya62kV40x2caYktHXAgBwBApXfDmdwkZZg5OiOYJ8mDGmStKZkl476qFSSXURt+tH7zuiIBtj7pB0hyRVVFRELScAILZRuABMhqifpGeMSZf0e0mfttb2HP3wMV5i33KHtfdZa2ustTUFBQXRiAkAAABIinJBNsZ4NVKOf22tfeQYT6mXVB5xu0xSQzQzAQAAACcSzVUsjKSfStpurb37OE97TNKto6tZnCOpm/nHAAAAcFI05yCfL+kWSZuNMRtG7/uipApJstbeK+lJSVdL2iNpQNKHo5gHAAAgqk7nRFKJk0ljRTRXsXhJx55jHPkcK+nOaGUAAACYTCy1lhjYSQ8A4hAbAgBA9EzKMm8AgInFKBUARA8jyAAAAEAECjIAAAAQgYIMAAAARKAgAwAAABEoyAAAAEAECjIAAAAQgYIMAAAARKAgAwAAABEoyAAAJLDT2XWRHRcxVbGTHgAACex0dl1kx0VMVYwgAwAAABEoyAAAAEAECjIAjNPpzO1kficAxC7mIAPAOJ3O3E6J+Z0AEKsYQQYAAAAiUJABAACACBRkAAAAIAIFGQAAAIhAQQYAAAAiUJABABgDtmwGpg6WeQMAYAzYshmYOhhBBgAAACJQkAEAAIAIFGQAAAAgAgUZAAAAiEBBBgAAACJQkAEAAIAIFGQAAAAgAgUZAAAAiEBBBgAAACJQkAEAAIAIFGQAAAAgAgUZAAAAiEBBBgAACae8olLGmHFdyisqnY4Ph3mcDgAAiA3WWtV3DmprQ492NvWqsXtQTT1Dau0d1nAwrGAorLCVij/0HT2yvl7JXreyU73KSU1SblqS8tKT5HEx7oLYUF9Xq7tX7hzXa+9aUT3BaRBvKMgAMEVZa7WjqVev7G3X6r1tWrO/Qz1DwcOP56cnqSgzWYUZPiV73fK4XXIZaceaDgXDVi29w9rT2idrR57vdhkVZfg0LTtFlXmpKslKkdtlHPrsAGD8KMgAMIWErdWhzkHlXHqHLvjW8zrUNShJqspL1dULS7SoLFtzSzJUXZyh1KRj/4i45+al+qePj4zMhcJW3YMBtfcPq6l7SA1dQ1pf26l1Bzvl87hUmZeq6flpqspLU7LXPWmfJwCcDgoyAEwBnf1+bWvs0Y6mXvUNB5W++ArNLcnUpy6dpQtm5Wtadsq43tftMspNG5liMaswQ5LkD4ZV2zGg/W392t/Wr13NfTJGmpaVooyz3q3a9gFV5KVO5KcHABOKggwACWo4GNKu5j5tb+xRY/eQjKTKvFRdOCtfP/rI+fqJfygqHzfJ49LMwnTNLEyXtVbNPcPa19anfW39yr3kdr3tv57XnOIMrZhXpBXzizV/WqaMYSoGgNhBQQaABNPSO6SNdd3a2dyrUNgqNy1JF8zM15ziDKX5Rr7t28DwpGQxxqg4K1nFWck674x8/eONb9MPH39FT29r1vee36N7ntujaVnJent1oS6cla/zzshTdmrSpGQDgOOhIAOIe+UVlaqvqx3Xa8vKK1RXe3CCE02+UNhqb2ufNtR1qbF7SB6X0dySDC2YlqXCDF/MjNAGu5t1+4UzdPuFM9TR79ez25u1cluzHt/YoAfX1MoYqbooQ8sqc7S0IkcLSrM0oyBNXjerYwCYPBRkAHFvKi/nNOAPavOhbm0+1K3+4ZCyUry6cFa+5pdkyhfjJ8XlpiXpxppy3VhTrkAorE31XXppd7vWHezQYxsa9OvXRv7T43UbnVGQrqq8NOVccrveqO1URrJXGckeZSR7lOJ1x8x/AAAkBgoyAMQhT1aRnt/Zom0NPQqGrSpzU3XpnGxV5aXGZVn0ul1aVpmrZZW5kv46Ir69sUfbG3u1s6lHe1r7lL74Sr24u+2I1xpJPq9LyR63kr1uJXtdSva65fO45HW75PW4lOR2KW3BpXpqc6PSfB6l+dxK842U66TRx5M8rsPX4/HvEMDEoSADQBzZcqhbP3pxn6bdcZ+2HOrWnOJMLavMUW5aYs3bdbuMZhdlaHZRhq5d8tf7jTH62hPb1DsUUO9QUL1DQQ36QxoMhDQcCGkoGNaAP6T2fr/8wbACo5ubSFL+Oz6jv/n1+jF9/KMLc5LHpZKP/kD/u6ZWHpeR22XkcRl5XC553Efe9rqNfKMF/c3i7s0rV0vPkDJTvCx3B8QBCjIAxDhrrV7Z2657X9irVbvblO7zqGfto7rrrs8pPXnqfRtP8bqV4nVrdFW5kwqGwwoErb7ygYu1eccu9Q8H1TccUv9wUEOBkPzBsPyhsIYDo38GwyP3BcPyh0KHr29fXaf0OXMVDIcVDFn5g2EFw0GFwvbwfSPX7VsyTLv9hzr7689Kknwel3LTklSY4VNBRrIKM30qzPCpMCNZJVnJKs9NVVlOCkUacNDU+84KAHEiFLZ6akujfvTCPm0+1K2CDJ8+f+UcvX95hbL/7Uqlf/ELTkeMCx6XS56kkRME5xRnjvt9vnvzUr3rb2876fPCYavhYFhDwZCGA2ENB0P68b99Wvf+7AH1DAZGNlbp86u1b1j1nQN6o7ZT7f3+t7xPcWayKnJTVZ6bqun5qZpTnKk5JRkqzU5hCggQZRRkAIgxwVBYj21s0Pee26N9bf2akZ+mb16/UO8+s5RRxTjgchmlJLmVkvTXYzWw/UXdck7lcV8TCIXV1jeshq4h1XUMqDbi8vKeNv1+/V/XrM5I9mjuaFmeV5KppZU5mlmQLhfbegMThoIMADEiGArr0Q0N+t5zu3WgfUBzSzL1ww8s1RXziyk/Cc7rdqkkK0UlWSlaVpnzlsf7hoPa2dSr7Y092tE0cuLi71+v1wP+kKSR0nxmRY6WVeRoaWW2lpRnKyPZO9mfBpAwKMgA4LBAKKw/vHFI339+jw62D2heSaZ+dMsyXT63iGIMSVK6z6NllTlHlOdw2OpAe7/W13bp9YOdeqO2U995dpes1eH1pJdW5ihtwSXqHPArO8XL1AxgjCjIAOCgJzY16tsrd2p/W78WlGbqx7fW6LK5hRQZnJTLZTSjIF0zCtJ1w7IySVLPUEAb60YK8+sHO/X4hgblv+MuPbD6oJI9LhVnJaskK0XFWckqyvTJ52HKDnAsFGQAcEBtx4CKb/1v3fm/6zW7KJ1iPBlM4q9vnJns1YWzCnThrAJJI6PMyUXTdct/PaymniE1dg/pQHv74efnpSWpJCtZRVnJKszwKS/NJze/tQAoyAAwmZp7hvTy3jbVdQzKnZqlb9+4WNedWUopmQw2PO4dF6X43HXR5TIKtB3UgtIsLSjNkiQNB0KHy3JTz5B2t/RpS0PPyPPNyA6HBRk+FaT75Cubr56hgDJjYD6ztVYH2wdGNo9p6tWupl419QyptXdYbX3DsnZk10WvxzWyhN51X9Kq3a0qzEhWaU6K0n1UHowd/1oAYBL0Dwe1ak+bdjb1Ktnr0oWz8vWrj1+nG3741uW9gGjyed2qzEtTZV6apJHi2TUYUGvv8Milb3i0iPaq+APf0qKvrlR5bormlWSquihDM4syNKswXdPz06K6qopJSlFyxUL5SqrlK50jX+lcudOyRzKHQwp2NijY06pQf5dCA11SOCTj8sh4vGpIz5MnZ5o21ncrFO6SJGWleFWVl6pZhRmalp2c8L9NwOmhIANAFIXCVhvru/Tavg6FwlY1lTmqqcqRz+PWr0IBp+MBMsYoJzVJOalJml30191X+oeD+tqnbtW3f/KQtjX2aHtDj57e1nx4Z0KXkSrz0jSzMF2zi9I1qzBDMwvTVZ6TqswUzykV0GAorK7BgJp7htQ0OrJd/qmHZFwjBTw7xauS7GSVZKaoMNOnvLQkedxzTvied62o1rf/tEOtfcM61DWo+s5BbWno0cb6bqUluTWzcCRzSXayXJRlHIWCDABRUtcxoL/salVHv1+Veal6++wC5aQm1pbQSFxpPo+G9r2uOy+eefi+4WBI+9v6tbu5T7tb+rSnpVe7mvv0/I6WI3YQTEtya1p2iopu/ob+uKlByV63XMbozR4aCI7sWDgUCKlnKKi+4eDh1/o8LhVlJqt79W91699+VkWZPqUmja+uuFxGRZnJKspM1tKKHPmD4ZH8Lb2Hy3JqkluzizI0tzhDBRk+RpYhiYIMABOubzioVbtataulT5nJHr1zUYmm56fxgxdxz+dxj+zod9SOhP5gWAfb+7Wnpe/waG1j96A2G6OugYCGAkMKW8nKylopyeOSz+NSsset8twUZaV4lZ0ysv12durIcnSrPv9rTf/ylyc0f5LHperiDFUXZ8gfDOtAe792Nfdqc323NtR1KS8tSXNLMuVOz53Qj4v4E7WCbIz5maRrJLVYaxcc4/GLJP2fpP2jdz1irf23aOUBgGiz1mpbY49e3N2mUNhq+fRc1VTmyON2OR0NiKokj0uzijI0K2KKhiTdd+tZ+vxpnBgZTUkel2YXZWh2UYaGAiHtau7VjqZevbSnTaV/83Pd8tPX9J6lZVoxv2jcI9iIX9E84vdL+p6kB07wnFXW2muimAEAJkXvUEDP7mjRwfYBTctO1uVzi5TNdAq86TSXmCsrr1Bd7cEJDIRIyV63FpVla1FZtjoH/Lrn29/UvtyP6NO/2aC0JLeuWlii9y+v0NKKt+5yiMQUtYJsrX3RGFMVrfcHgFixp6VPz2xvVihsddHsAi0qy2I6BY40BZeYi1c5qUnqfunXWvXCL7XmQIceWV+vJzc36Xev12v59Fz9zUVn6O2zC/gaT3BO/87gXGPMRkkNkj5nrd3qcB4AGDPj9enZ7c3a0tCjwgyfrlxQzEl4QIJwuYzOmZGnc2bk6SvvnK+H1tbpJ6v26bafr1VNZY6+fv3CI1b9QGJxcmLcekmV1trFkv5H0qPHe6Ix5g5jzDpjzLrW1tbJygcAx7W/rV/Ft9ytLQ09qqnM0XtryinHQIJK83n00Qum64V/uFjfuH6h9rb26R33rNLdK3dqKBByOh6iwLGCbK3tsdb2jV5/UpLXGJN/nOfeZ62tsdbWFBQUTGpOADjaM9ua9a7/eUnutGy9e8k0nT8zn53wgCkgyePSzWdX6Jm73q5rFk3TPc/t0Q33vqKWniGno2GCOVaQjTHFZnQCjzHm7NEs7Sd+FQA4x1qr7z6zW7c/sE5V+Wlq/MWnD+9GBmDqyEv36b/ft0Q/vrVG+1r7dd0PXtGell6nY2ECRa0gG2MelLRaUrUxpt4Y81FjzCeMMZ8YfcoNkraMzkG+R9JN1lp7vPcDACf5g2F99rcb9d/P7NJ7lpbp4U+cq1APU76AqezyeUX6zR3najgY1nt+uFrrDnQ4HQkTJJqrWNx8kse/p5Fl4AAgpnUPBPTxX63Tq/s69LkVs3XnxTM5gx2AJGlhWZb+8Lfn6UM/W6MP379WntwypyNhArB6PQCcQEvPkN77o9Vaf7BL371pif7uklmUYwBHKM9N1QMfPVs+j0uFN3xZg5y4F/coyABwHPWdA3rvj1arrnNA93/4LF27pNTpSABiVFlOqn50yzJ5Mgr05KZGhcLMGo1nFGQAOIb9bf16772r1d7v1y8/ulznzTzmIjsAcNiyyly1/+ke1XcNatVuzlGIZxRkADjKwfZ+3XTfag0Fw3rwY+doWSXbywIYm/6tz2tJebY21nervnPA6TgYJwoyAERo6BrU+3/8moaDYf3vx5ZrQWmW05EAxJnzzshTVopXz2xvUSAUdjoOxoGCDACjWnqH9IGfvKaewYB++ZHlmlOc6XQkAHHI63bp0jmF6h4M6LV9LP0WjyjIACCpdyigW3+6Rs09Q7r/I2dpYRkjxwDGrzw3VQumZWp9baea2Wkv7lCQAUx5/mBYf/Or9drT0qd7P7hMyypznY4EIAFcMDNfqUluPbejReyFFl8oyACmvC88skkv7WnTN65fqLfNLnA6DoAE4fO6df7MfLX0DmtPS5/TcXAKKMgAprSs89+vR9Yf0qcvm6Uba8qdjgMgwVQXZygvLUmv7G1nbeQ4QkEGMGXtbulV9gXv13uWlulTl85yOg6ABOQyRuedkaeuwYC2NfY4HQdjREEGMCW19Q3r6W3NGm7Yoa9fv4DtowFEzfT8NJVkJeu1/e0s+xYnKMgAppzBQEiPb2xQktul1j98XT6P2+lIwIkZl4wx47rAecYYnX9GvvqHQ9pY1+V0HIyBx+kAADCZwtbqT1ua1D8c0g3LyrStjzVKEQdsWHev3Dmul961onqCw2A8SnNSVJmXqtdrO7W4PFteN2OUsYyjA2BKWXegU7UdA7qoukDFWclOxwEwhZxVlauhQFhbG5iLHOsoyACmjPrOAb26r13VxRmaP41d8gBMrtLsFJVkJWt9bScrWsQ4CjKAKaF/OKintjQpO9WrS6oLmZsJwBE1lTnqHQpqd0uv01FwAhRkAI4rr6gc9wlIYym61lqt3Nas4WBYVy0oUZKHb32c9AU4Y3p+mnLTkvT6wU5214thnKQHwHH1dbXjPgFJOvlJSBvqulTbMaCLqwtUkOEb98dJKJz0BTjCGKNlFTl6enuzDrYPqCo/zelIOAaGUQAktLa+Yb28t13T89O0sDTL6TgAoOriDKX7PHr9YKfTUXAcFGQACSsYDuvPW5uU5HbpsrnMOwYQG9wuoyXl2arvGlRr77DTcXAMFGQACWv13na19fl1+bwipSYxowxA7Jg/LVMel9HG+i6no+AYKMgAElJD16DW13ZpQWmmpjPHD0CMSfa6Nac4QzuaejUYCDkdB0ehIANIOIFQWE9va1ZmskcXzixwOg4AHNPi8myFwlZbD3U7HQVHoSADSDir97WrazCgy+YWsaQbgJiVn+5TWU6KNtZ3K8zGITGFnxwAEkpD16DeqO3SwtIsleemOh0HAE5oSXm2+oaD2tvW53QURKAgA0gYwVBYT28fmVpxwcx8p+MAwElNz09TRrJHG+uYZhFLKMgAEsbaA53qGgjokjmFTK0AEBdcxmhxWbYOseRbTBnTTxBjzO+NMe8wxvATB0BMausb1rqDHZpTnKHKPFatABA/WPIt9oy18P5Q0vsl7TbGfNMYMyeKmQDg1BiXnt3eIp/HrbfNYtUKAKfJuGSMGddlPJK9bs0pGVnyzZWSOcGfDMZjTCvnW2ufkfSMMSZL0s2SnjbG1En6saRfWWsDUcwIACeUcebVauoZ0hXzipSS5HY6DoB4Z8O6e+XOcb30rhXV43rdkrJsbTnUo/RFK8b1ekysMU+ZMMbkSbpN0u2S3pD0XUlLJT0dlWQAMAb9w0Flv+1WVeSmqro4w+k4ADAueaNLvmUsvVrBUNjpOFPeWOcgPyJplaRUSe+01r7LWvsba+3fS0qPZkAAOJFVe9pk3F5dVF0w7l9vAkAsWFKeLU9moZ7e1ux0lClvrCPIP7HWzrPWfsNa2yhJxhifJFlra6KWDgBOoL5zQDubetW95vfKSU1yOg4AnJbp+WkKdjXp568ccDrKlDfWgvwfx7hv9UQGAYBTEQpb/WVnqzKTPepZ/bDTcQDgtLmMUc/6J7Rmf4e2NrAuspNOWJCNMcXGmGWSUowxZxpjlo5eLtLIdAsAcMTGui619/v19tkFskHWDgWQGPo3rVSK161fMIrsqJOtYnGFRk7MK5N0d8T9vZK+GKVMAHBCfUNBvbq/XdPz0zSjgNMgACSO8HC/rltaqt+9Xq8vXDVXuWlMH3PCCUeQrbW/sNZeLOk2a+3FEZd3WWsfmaSMAHCEVbtbFbbS22ez5jGAxHPbeVXyB8N6aG2t01GmrBOOIBtjPmit/ZWkKmPMXUc/bq29+xgvA4Coqe0Y0K6WPi2fnqusFK/TcQBgws0uytD5M/P0y9UHdceFM+Rxs5HxZDvZ3/ib+7WmS8o4xgUAJs3IiXktykrxqqYyx+k4ABA1t503XY3dQ1rJkm+OOOEIsrX2R6N//uvkxAGA43ujrlOdAwG9a/E0RlQAJLRL5hSqPDdFP395v65eWOJ0nClnrBuF/KcxJtMY4zXGPGuMaTPGfDDa4QDgTf3DQa3Z36EZ+Wmanp928hcAQBxzu4xuO2+61h7o1PraTqfjTDljHYJZYa3tkXSNpHpJsyX9Q9RSAcBRXtnbrnBYunBWvtNRAGBS3HRWubJSvLrvhX1OR5lyxlqQ3zwT5mpJD1prO6KUBwDeorlnSNsae7SkPFvZ7JgHYIpI83l0yzmV+vO2Ju1r7XM6zpQy1oL8uDFmh6QaSc8aYwokDUUvFgCMsNbqxd2tSvG6ddZ0TswDMLV86Lwqed0u/XjVfqejTCljKsjW2i9IOldSjbU2IKlf0rXRDAYAkrSnpU8NXUM694w8+Txup+MAwKQqyPDphmVl+v36erX0MjY5WU7lNPC5kt5njLlV0g2SVkQnEgCMCIbCemlPm/LTkzR/WqbTcQDAER+7cIYCoTDbT0+isa5i8UtJ35Z0gaSzRi81UcwFANpQ16WeoaAunFUglzFOx5l4xiVjzLguAKaO6flpunJ+sR5YfVDdgwGn40wJJ1wHOUKNpHnWWhvNMADwpv7hoNYe6NSM/DRV5KY6HSc6bFh3r9w5rpfetaJ6gsMAiGV3XjxTT21p0v0vH9CnLpvldJyEN9YpFlskFUczCABEWr2vXcFwWBewrFtiOY1Rc0bOMZUtKM3SinlF+ulL+xhFngRjHUHOl7TNGLNG0vCbd1pr3xWVVACmtNbeYW1t6NGZFdnKifaybqOFDZPkNEbNJUbOMbV98tJZWrmtmVHkSTDWgvzVaIYAgDdZa/XirlYle11aXpU7CR+QaQ4A4kPkKPJt51cpK8V78hdhXMa6zNsLkg5I8o5eXytpfRRzAZii9rX1q75rUOfMyJPPy7JuABDpk5fOUs9QUPe/fMDpKAltrKtYfEzS7yT9aPSuUkmPRikTgCkqGA5r1e425aYlaeG0LKfjAEDMeXMU+Scv7VNnv9/pOAlrrCfp3SnpfEk9kmSt3S2pMFqhAExNG+u61T0Y0Ntm5cvlYl4wABzLZ1dUq384qO8/v8fpKAlrrAV52Fp7+L8pxhiPJJZ8AzBhBvxBrdnfoaq8VFXmpTkdBwBiVnVxhm5cVq4HVh9UXceA03ES0lgL8gvGmC9KSjHGXC7pYUmPRy8WgKnm1X0dCoTDunBWgdNRACDmfeby2XK5pP/68/hXhcHxjbUgf0FSq6TNkj4u6UlJ/xytUACmFm9+pbYc6tbi0mzlpkV5WTcASADFWcm6/YIZemxjgzbVdzkdJ+GMdRWLsEZOyvtba+0N1tofs6segIlgrVXOpbcryePS8hmTsKwbACSIj799hvLSkvT1J7eLWjaxTliQzYivGmPaJO2QtNMY02qM+fLJ3tgY8zNjTIsxZssJ3vseY8weY8wmY8zS8X0KAOLZcztalFJ1ps6ZkadklnUDgDHLSPbqU5fN0qv7OvSXna1Ox0koJxtB/rRGVq84y1qbZ63NlbRc0vnGmM+c5LX3S7ryBI9fJWnW6OUOST8cS2AAicMfDOtrT2xXoL1OC0tZ1g0ATtXNZ1doen6avvHUdsmMdeYsTuZkf5O3SrrZWrv/zTustfskfXD0seOy1r4oqeMET7lW0gN2xKuSso0xJWOLDSAR/PLVg9rX1q+O534qN8u6AcAp87pd+scrqrWruU9pCy51Ok7COFlB9lpr246+01rbKul09zcslVQXcbt+9L63MMbcYYxZZ4xZ19rKrxCARNDR79d3n9mlC2fla2jfOqfjAEDcunJBsc6syFb2hR9UIBR2Ok5COFlBPtEWLae7fcuxhouOOcPcWnuftbbGWltTUMASUEAi+M4zu9Q3HNS/XDPP6SgAENeMMfrS1XPlycjTG3VdTsdJCJ6TPL7YGNNzjPuNpOTT/Nj1ksojbpdJajjN9wQQB3Y39+rXr9XqA8srNbsow+k4ABD3aqpyNbDzFb3uvkALpmUqNelkFQ8ncsIRZGut21qbeYxLhrX2dKdYPCbp1tHVLM6R1G2tbTzN9wQQB/7jie1KTXLrM5fPdjoKACSMzhd/oUA4rNf2n+gUMIxF1P57YYx5UNJFkvKNMfWSvqLRecvW2ns1stnI1ZL2SBqQ9OFoZQEQO57f2aIXdrXqn98xl01BgFhnXDKGE2jjRbDjkBZOy9KWQ91aUp6tnFS+x45X1AqytfbmkzxuJd0ZrY8PIPYEQmH9xx+3qSovVbeeW+V0HAAnY8O6e+X4tzK+a0X1BIbBWJw9PVfbm3r08p42XbNomtNx4hYL5gGYNA+sPqi9rf360jvmKcnDtx8AmGhpPo9qKnO1t7Vfh7oGnY4Tt/gJBWBStPcN6zujy7pdNrfQ6TgAkLDOrMhWWpJbL+1uYwvqcaIgA5gU3165U4P+kL7yznnMaQSAKPK6XTpnRp6aeoa0v63f6ThxiYIMIOq2HOrWQ2vr9KHzqjSzkGXdACDa5pZkKivFq9X72hlFHgcKMoCostbqq49tVW5qkj556Syn4wDAlOB2GZ0zPVdtfX7taelzOk7coSADiKrHNjZo3cFO/cMV1cpKOd3l0wEAYzW7OEO5qUl6dV+HwowinxIKMoCoGfAH9Y0nd2hBaaZurCk/+QsAABPGZYzOmZGrjgG/djX1Oh0nrlCQAUTND57fq6aeIX31nfPldnFiHgBMtpmF6SpI9+nV/R0KhxlFHisKMoCoqG0f0H2r9undS6appirX6TgAMCUZY7R8Rq66BwPazVzkMaMgA4iKrz25TR6X0Reumut0FACY0mbkpyk3NUnrDnawosUYUZABTLjnd7boz1ubdefFM1Wclex0HACY0owxWlaZo7Y+vw62DzgdJy5QkAFMqEF/SP/y6BadUZCm2y+c7nQcAICk6uIMpfs8Wnew0+kocYGCDGBC3fPcbtV3Dupr1y2Uz+N2Og4AQCPrIi+tyNahrkE1dg86HSfmUZABTJhdzb368Yv7dMOyMp0zI8/pOACACPOnZSnZ49K6A4winwwFGcCECIetvvSHzcpI9uiLV3NiHgDEmiSPS4vKs7WvrV+dA36n48Q0CjKACfHw63Vae6BT/3T1XOWmJTkdBwBwDItKs+Qy0qb6bqejxDQKMoDT1t43rG88tUNnV+XqxmVlTscBABxHms+jmYXp2tbYI38w7HScmEVBBiBJKq+olDFmXJfq9/2TOnsH9bXrFsgYdswDgFi2uCxb/mBYO9l++rg8TgcAEBvq62p198qdp/y6uo4BPfLGIXW/8hvNKro2CskAABOpJCtZBRk+bazv0oLSTAY2joERZADjFgyF9fzOFmWleNW9+jdOxwEAjIExRovLstTe79ehLpZ8OxYKMoBxe21/hzoHArq4ukA2yBnRABAvqosylOxxaWMdJ+sdCwUZwLg09wzp9dpOzSvJVGVemtNxAACnwON2aX5plva29alvOOh0nJhDQQZwyoLhsJ7e1qzUJLfeNivf6TgAgHFYMC1T1krbG3ucjhJzKMgATtna/Z1q7/fr0jlF8nnZThoA4lF2apJKs1O0taFH1lqn48QUCjKAU9LcM6R1Bzs0pzhD0/OZWgEA8WzBtEx1DwY4We8oFGQAYxYMhfXnrU1KTfLo7bMLnI4DADhNMwvTleRxaWsD0ywiUZABjNnLe9rVORDQZXMLlczUCgCIex63S9VFGdrd0ifj47eCb6IgAxiTg+392lDfpSVl2axaAQAJZMG0TIXCVmnz3u50lJjBTnoATmooENIz21uUk+rV+TPznI4DAJhAhZnJKkj3aXjRCqejxAxGkAGckLVWz2xv1oA/qCvmF8vj5tsGACSa+dMy5SueqW3MRZZEQQZwElsO9Whva7/OOyNfRZnJTscBAETB7KIM2VBAf3ij3ukoMYGCDOC42vuG9cLuVlXkpmppRbbTcQAAUZKS5Nbg3nV6dEODgqGw03EcR0EGcEyBUFhPbWlSktulFfOKZIxxOhIAIIr6tj6n1t5hvby33ekojqMgA3gLa62e39mi9n6/VswvUpqP83kBINEN7l2rrBSvHlnPNAsKMoC32NrYo+2NvTq7KldVLOkGAFNDKKhrFpXoz1ub1DccdDqNoyjIAI7Q0jukv+xsVXluipbPyHU6DgBgEl2/tExDgbCe2tzodBRHUZABHDYcCOnJzU1K8bp15fxiuZh3DABTytKKbFXlpeoPbxxyOoqjKMgARhmt3Nas3qGArl5YrNQk5h0DwFRjjNF1Z5Zp9b52HeoadDqOYyjIACRJmcuv1762fl0wM18lWSlOxwEAOOS6M0tlrfToFB5FpiAD0Oq97cp+262aVZiuJeXZTscBADioIi9VZ1fl6g9vHJK11uk4jqAgA1Ncc8+Q/v7BNxTsbNBlc1nvGAAgXbe0VHta+rT5ULfTURxBQQamsKFASHf88nUN+INqffQbSvLwLQEAIF29sERJHpceWT81p1nw0xCYoqy1+udHt2hjXZfufu8SBdpqnY4EAIgRWSleXT63SI9tbFBgCm49TUEGpqj7Xzmg371er09eOktXLig+/Tc0LhljxnUBAMSe65eWqqPfrxd2tjodZdKxjhMwBb2yp03/8cR2XT6vSJ++dNbEvKkN6+6VO8f10rtWVE9MBgDAhHnb7ALlpSXpkTfqddm8IqfjTCpGkIEppq5jQHf+73pNz0/T3e9dLJeLEVwAwFt53S69c/E0PbO9Rd0DAafjTCoKMjCFDPiD+tgD6xQKW/341hplJHudjgQAiGHvWVomfzCsJ6bY1tMUZGCKsNbqHx7epF3Nvbrn5jM1PT/N6UgAgBi3oDRTswrT9cj6eqejTCoKMjBF/OAve/XE5kb945VzdFF1odNxAABxwBij65aWat3BTh1s73c6zqShIANTwHM7mvXtlTv1rsXT9PG3zXA6DgAgjrx7SamMkf4whbaepiADCW5PS58+9eAGzSvJ1Lfes4hl1QAAp2RadorOnZF3wq2nyysqx73UpzFG5RWVk/xZnRjLvAEJrGcooDt+uU5JHpfuu7VGKUlupyMBAOLQ9UvL9LmHN2p9baeWVea+5fH6utpxL/Upxd5yn4wgAwkqHLb6zEMbVNs+oO9/YKlKs1OcjgQAiFNXLihWitet30+RracpyECC+u9ndunZHS36yjvn6ZwZeU7HAQDEsXSfR1fML9IfNzZoKBByOk7UUZCBBPTU5kb9z3N79L6acn3wnNia1wUAiE/vWVamnqGgVm5rdjpK1FGQgQSzs6lXn314o86syNa/vXs+J+UBACbE+WfkqywnRQ+tqXU6StRRkIEE0jXg18ceWKc0n0f3fnCZfB5OygMATAyXy+h9NeV6ZW97wq+JTEEGEkQobPX3D76hxu5B3fvBZSrKTHY6EgAgwdxYUy6XkR5aW+d0lKiiIAMJ4j//vEOrdrfp369doGWVOU7HAQAkoOKsZF0yp1C/e71egVDY6ThRE9WCbIy50hiz0xizxxjzhWM8fpExptsYs2H08uVo5gFi3XgXWk+b93b96IV9Cu96UTedXeH0pwEASGA3nVWh1t5hPbejxekoURO1jUKMMW5J35d0uaR6SWuNMY9Za7cd9dRV1tpropUDiCfjWWi9tXdYv11Xp8IMn1b/392SvhWdcAAASLqoukBFmT49tKZWV8wvdjpOVERzBPlsSXustfustX5JD0m6NoofD5hyBvxBPb6pQclet65eWCKFg05HAgAkOI/bpRuXleuFXa2q7xxwOk5URLMgl0qKnMFdP3rf0c41xmw0xjxljJl/rDcyxtxhjFlnjFnX2toajaxA3AmFrZ7a0qQBf0jXLCpRms8jGde4pmiwFBwA4FTcvLxCxhj98tWDTkeJiqhNsZB0rJ+49qjb6yVVWmv7jDFXS3pU0qy3vMja+yTdJ0k1NTVHvwcwJb20p031nYNaMa/orytW2PApT9F4010rqicwHQAgkZVmp+iK+UV6aE2dPnXpW6pb3IvmCHK9pPKI22WSGiKfYK3tsdb2jV5/UpLXGJMfxUxAQtjW2KMNdV1aUp6tuSWZTscBAExBHz5/uroHA3pk/SGno0y4aBbktZJmGWOmG2OSJN0k6bHIJxhjis3o73aNMWeP5mmPYiYg7jX1DOm5HS0qy0nRhTP5/yQAwBk1lTlaUJqp+1854HSUCRe1gmytDUr6O0l/lrRd0m+ttVuNMZ8wxnxi9Gk3SNpijNko6R5JN1lrmUIBHEf/cFBPbGpUapJbVy8okcvF3GEAgDOMMfrI+dO1p6VPyVVnOh1nQkVzDvKb0yaePOq+eyOuf0/S96KZAUgUobDVk5sbNRQI6b015UpJYhtpAICz3rGoRF9/cocGat7ldJQJxU56QJx4YVerGrqHdNncIhVk+JyOAwCAfB63PnhOhVLPOEsd/X6n40wYCjIQB7Yc6tbmQ91aVpmj6uIMp+MAAHDYredWKewf1JoDHU5HmTAUZCDGNXQN6vmdLarITdV5Z+Q5HQcAgCPkpiWp940ntaupV50DiTGKTEEGYljfcFBPbG5URrJXVy0olosNPQAAMahnzSNyu4zW7E+MUWQKMhCjguGwntjUqEAorHcuKlGyl5PyAACxKTzQrYVlWdrZ1KuuBBhFpiADMchaq+d3tKqpZ0gr5hUrL52T8gAAsW1ZRY5cLpMQc5EpyEAM2lTfrW2NPTq7KlczC9OdjgMAwEml+TxaWJqlHQkwF5mCDMSY+s4Bvbi7VdPz03TOjFyn4wAAMGY1lTnyuIxe2t3mdJTTQkEGYog7o0BPbm5SVopXV8wvkuGkPABAHEnzeXRWVa72tfWrtmPA6TjjRkEGYsRQIKSC67+kUNjqnYumyefhpDwAQPw5szxbmckevbirVeGwdTrOuFCQgRhgrdU//G6Tkopm6Ir5RcpJS3I6EgAA4+Jxu3ThrAK19/u1uaHb6TjjQkEGYsB3n92txzc2qOuFX2hGASflAQDi2xkFaSrLTtGr+9o1FAg5HeeUUZABhz2+sUHfeWa33rO0TD2v/d7pOAAAnDZjjN42u0D+YFgv7Gp1Os4poyADDtpQ16XPPbxRZ1fl6uvXL3A6DgAAE6Ygw6eaqlztaOrVvtY+p+OcEgoy4JCGrkHd/ot1Ksz06d5blnFSHgAg4Zxdlav89CQ9u6MlrqZaUJABB/QPB/XRX6zTcCCkn33oLOVyUh4AIAG5XUYr5hVrKBCKq6kWFGRgkoXDVp/+zQbtbOrR/7z/TM0qynA6EgAAURM51WJXc6/TccaEggxMsq89uV1Pb2vWl6+Zp4uqC52OAwBA1J1dlauSrGQ9u70lLrahpiADk+jHL+7TT1/ar9vOq9KHzqtyOg4AAJPC7TK6akGxXC7pic2NCoTCTkc6IQoyMEn+b8Mhfe3J7XrHwhJ9+Zp5bCMNAJhSMpK9umJ+sdr7/PrLztiej0xBBibBy3va9LmHN2r59Fz9v/culstFOQYATD1VeWk6uypX2xp7tKm+y+k4x0VBBqJsa0O3Pv7L13VGQbruu7VGyV6WcwMATF3LZ+SqKi9Vf9nVqtqOAafjHBMFGYiiuo4B3fbztcpM9uj+D5+trBSv05EAAHCUyxhduaBYualJenJzY0yetEdBBqKko9+vD/1sjfzBsH7xkbNVnJXsdCQAAGKCz+PWOxdPk8sYPbahQS5fmtORjkBBBqKgeyCgW376mg51DeonH6phrWMAAI6SleLVOxaVqN8fVFLpXKfjHIGCDEywnqGAbv3Za9rd3Kcf3bJMZ1XlOh0JAICYVJqdog+fP11D+9Y5HeUIFGRgAvUNB3Xbz9Zoa0OPfvCBpWwEAgDASaTE4MnrHqcDAIliwB/UR36+Vhvru/X995+py+YVOR0JAACMAyPIwAQY9If00fvXad3BDn33piW6ckGJ05EAAMA4MYIMnKahQEh3/HKdXt3frv9+7xJds2ia05EAAMBpYAQZOA0D/qBu/8U6rdrdpv98zyK9+8xSlVdUyhgzrgsAAHAeI8jAOHUPBvThn6/Rhrou/dcNi3RjTbkkqb6uVnev3Dmu97xrRfVERgQAAONAQQbGoa1vWLf+dI12t/Tq++9fqqsWMucYAIBEwRQL4BhONE3CmzNNiz/7gLbUtqj+wX/R1YumMU0CAIAEwggycAzHmybR1D2kxzY2yMrqXYunqeTKB9/yHKZJAAAQ3yjIwBjtbe3Tn7Y0Kc3n0bVLpiknNcnpSAAAIAooyMBJWGu19mCnVu9tV1GmT+9cNE1pPr50AABIVPyUB04gGArr6e3N2tXcp+qiDF02t1AeN1P3AQBIZBRk4Di6Bvx6cnOTWvuGdd4ZeaqpzOEkPAAApgAKMnAMKbPO1YNr6mSM9K7F0zQ9P83pSAAAYJJQkIEIQ4GQvvWnHSq8/kvKSfPq6gUlykzxOh0LAABMIiZTAqO2N/bo2u+9rJ+/fEA96x7TDcvKKMcAAExBFGRMecFQWPe+sFfXfu9ldQz4df+Hz1Lns/fJ4+LLAwCAqYgpFpjStjZ06/O/36Qth3p0xfwiff26hcpL9zkdCwAAOIiCjCmpbzio/3lut36yar9yUpP0gw8s1VULilmlAgAAUJAxtVhr9X8bGvT1J7erpXdYNy4r05feMVfZ7IoHAABGUZAxZazZ36FvPLVdb9R2aVFZlu69ZZmWVuQ4HQsAAMQYCjIS3vbGHn37zzv17I4WFWX69K33LNSNy8rlcjGdAgAAvBUFGQmpvKJSLX6vss67Samzz1V4qE/drz6s2tf/qJu+NKybnA4IAABiFgUZCcVaq5f3tMu//CMqOaNGPo9LS8qztaR8hpLfcaakr4/pfe5aUR3doAAAIGZRkJEQ+oaD+uPGBt3/ygHtaOpVUtEZOveMPC0uy5LP43Y6HgAAiCMUZMQta63W13bpt2vr9PimBg34Q6ouytB/3rBI7ztnhs5+aqvTEQEAQByiICPutPUN69E3Duk3a+u0u6VPqUluXbOoRO87q1xLK3JkjNH7QkGnYwIAgDhFQUZcaOoe0p+3NumpLY1as79DYSudWZGtb16/UNcsnqZ0H/+UAQDAxKBVICZZa3WwfUBPb2vWU1satb62S5I0szBdd148U9csmqbq4gxnQwIAgIREQUbM6Oz36+W9bXp5T5tW7W5TfeegJGleSaY+e/lsXbWwWDMLKcUAACC6KMhjVF5Rqfq62nG/vqy8QnW1BycwUXyz1qqxe0gb6rr0Rm2nVu9r19aGHlkrZfg8OueMPNWu/Lka1z+jg11NesrpwAAAYMqgII9RfV2t7l65c9yvn8rr6lpr1do7rF3Nfdp0qEsbaru0oa5LLb3DkqQk98haxZ+5bLbOn5mvxWVZ8rhdMh/61bj/zqfy3zcAADg9FOQx6B0KyJ2eqwF/UG5j5HIZuYyRy0jGsF3xmwb9IR3qGlB956D2tvZrT0uvdjf3aXdLn7oHA4efV5WXqvPOyBvZwKMiR3NLMlirGAAAxIyoFmRjzJWSvivJLekn1tpvHvW4GX38akkDkm6z1q6PZqbx+PnLB1R25wP68ar9b3nMZaQkj0s+j1tJbpeSPH+9+CJupy+5So9tbFBmskeZKV5lJnuVmeJRZrJXyd7YLYfBUFj9wyF1DPjV0T+stj6/OvpHLm19w2rsGtKhrkEd6hpUR7//iNdmp3o1uzBD1ywq0azCdM0qytC8kkzlpCU59NkAABDjjIvBtxgQtYJsjHFL+r6kyyXVS1prjHnMWrst4mlXSZo1elku6Yejf8aUS+YU6kuf/aTe88l/VchaWSuFrFU4bBUKW/mDYflDYQ0Hw/IHw+obCmo4GJI/NHI7bKW8K+7UJx9845jvn+RxHVGYM5I9Sva6R4v3SPn2eVzyeSOue478Aor8Uor8ugqGrPyhsIIhq0AorEA4rEBw5HowHJY/aDXgD6rfH1L/cHDk4g+qfzikvuGg/MHwcf9e0n0eFWcla1p2ihaUZqksJ0Wl2Smalp2i6flpWjp/th4+jXnbAABMOTbs3PRCyvlh0RxBPlvSHmvtPkkyxjwk6VpJkQX5WkkPWGutpFeNMdnGmBJrbWMUc52yBaVZ6tv4Jy0u/+4pv9Zaq2DY6os3nKcd+w6qezConqGAegYD6hkKjv4ZUE/E/b1DQbX1+TUcDGk4MFK8h4OhwwV8vKwNywYDUjgkGwrIhoJSOKRwYFDWP6Swf1DWP3jEnwr5FRrqU3iwV6GBboUHuhUavSgU0Mn2qmPeNgAAccLJch5jzEg3jcIbG3ODpCuttbeP3r5F0nJr7d9FPOePkr5prX1p9Pazkj5vrV131HvdIemO0ZvVksbfusYvX1KbAx8Xk4djnNg4vomN45vYOL6Jz6lj3GatvfLoO6M5gnysMfqj2/hYniNr7X2S7puIUONljFlnra1xMgOii2Oc2Di+iY3jm9g4vokv1o6xK4rvXS+pPOJ2maSGcTwHAAAAmDTRLMhrJc0yxkw3xiRJuknSY0c95zFJt5oR50jqjrX5xwAAAJhaojbFwlobNMb8naQ/a2SZt59Za7caYz4x+vi9kp7UyBJvezSyzNuHo5VnAjg6xQOTgmOc2Di+iY3jm9g4vokvpo5x1E7SAwAAAOJRNKdYAAAAAHGHggwAAABEoCAfxRhzpTFmpzFmjzHmC8d43Bhj7hl9fJMxZqkTOTE+Yzi+Hxg9rpuMMa8YYxY7kRPjc7LjG/G8s4wxodH12hFHxnKMjTEXGWM2GGO2GmNemOyMGL8xfI/OMsY8bozZOHp8Y/ncJRzFGPMzY0yLMWbLcR6PmY5FQY4QsT32VZLmSbrZGDPvqKdFbo99h0a2x0YcGOPx3S/p7dbaRZL+XTF20gCOb4zH983nfUsjJxAjjozlGBtjsiX9QNK7rLXzJd042TkxPmP8Gr5T0jZr7WJJF0n6f6MrZSE+3C/pLZtyRIiZjkVBPtLh7bGttX5Jb26PHenw9tjW2lclZRtjSiY7KMblpMfXWvuKtbZz9OarGlmbG/FhLF+/kvT3kn4vqWUyw2FCjOUYv1/SI9baWkmy1nKc48dYjq+VlGGMMZLSJXVICk5uTIyXtfZFjRyz44mZjkVBPlKppLqI2/Wj953qcxCbTvXYfVTSU1FNhIl00uNrjCmVdJ2keycxFybOWL6GZ0vKMcb8xRjzujHm1klLh9M1luP7PUlzNbKp2GZJn7LWhicnHiZBzHSsaG41HY8mbHtsxKQxHztjzMUaKcgXRDURJtJYju93JH3eWhsaGYBCnBnLMfZIWibpUkkpklYbY1611u6KdjictrEc3yskbZB0iaQzJD1tjFllre2JcjZMjpjpWBTkI7E9dmIb07EzxiyS9BNJV1lr2ycpG07fWI5vjaSHRstxvqSrjTFBa+2jk5IQp2us36PbrLX9kvqNMS9KWiyJghz7xnJ8Pyzpm3ZkE4c9xpj9kuZIWjM5ERFlMdOxmGJxJLbHTmwnPb7GmApJj0i6hRGnuHPS42utnW6trbLWVkn6naS/pRzHlbF8j/4/SRcaYzzGmFRJyyVtn+ScGJ+xHN9ajfx2QMaYIknVkvZNakpEU8x0LEaQIyTg9tiIMMbj+2VJeZJ+MDrKGLTW1jiVGWM3xuOLODaWY2yt3W6M+ZOkTZLCkn5irT3mklKILWP8Gv53SfcbYzZr5Nfxn7fWtjkWGqfEGPOgRlYfyTfG1Ev6iiSvFHsdi62mAQAAgAhMsQAAAAAiUJABAACACBRkAAAAIAIFGQAAAIhAQQYAAAAiUJABAACACBRkAAAAIML/B1ORFdX+DpS9AAAAAElFTkSuQmCC\n",
      "text/plain": [
       "<Figure size 720x360 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 720x360 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAsgAAAFkCAYAAAA9nc1+AAAAOXRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjUuMSwgaHR0cHM6Ly9tYXRwbG90bGliLm9yZy/YYfK9AAAACXBIWXMAAAsTAAALEwEAmpwYAABFDUlEQVR4nO3dd3yV5cH/8c91TvYmi+ywNzJkiijYqjjQat0DF+JsVfy19Wl9tOt5nrZW3KPWiXvW0VLFjSIr7CGBACEJK5ssMs/1+yOxPcUAgeTkPkm+79frvMg5587JN95CvrnOdV+XsdYiIiIiIiLNXE4HEBERERHxJyrIIiIiIiJeVJBFRERERLyoIIuIiIiIeFFBFhERERHxooIsIiIiIuIlwOkAIiI9nTFmGjCt5e6D1tpyp7KIiIgKsoiIP5gG3Nvy8fNAuVNBREREUyxERERERP6DCrKIiIOMMV/w79FjgB3GGGuMyW15fogx5lVjzF5jTL0xZqcxZp4xJtLrNaa1fI41xtxsjHnaGFNpjMkzxlxtjHEbY35vjCkyxuwxxtxnjAnw+vznvT5/pDHmS2PMAWNMrjHm5s76byEi4i+MtpoWEXFOS0E+uZWndgLnAl8DEa08vxKYYq2ta5nD/HnL46VArNdxFvg7MPOgz7/JWvtkS4bngataHi8B4g469nJr7StH/m5ERLoHjSCLiDjIWjsN+I3XQ32ttcZa2weYR3M53goMB0KAH7ccdzxwTSsvWQMM5N+F2ABnATOAvkBVy+MXHCLSxzQX7JOA6pbHfnc035OISFengiwi4oeMMWH8e2R5ILARqAXe9jpseiuf+py1Ngf4xOuxxdbaj6y1ucD6lsfSD/Gl77bWlllrvwLeaXmsnzEm6Ri+DRGRLkkFWUTEP/UC3Ec4JraVx/IArLW1Bz/Wor7lz+BDvGa+18e7vD5OPUIWEZFuQwVZRMR5rV0MUgZ4Wj5+t2XaxX/cgNNa+bzGNj52KGleH3uX4l0HHygi0l2pIIuIOK/M6+MRANbaGmBRy2MzjTFXGmPCjTExxpizjTEf0zxPuKP9zhjTyxgzFTi/5bFt1tq9PvhaIiJ+SQVZRMR5K7w+/qBlubWXgDtpvlDODcyn+QK7MuAD4Ic0X4DX0U6heSWMRUB4y2P3+ODriIj4LRVkERGHWWuXAL+kef6vx+vxVcB44FVgH9AA7KF5SbfbgVU+iDO95fVraZ67fKuWeBORnkbrIIuI9HDe6yC3zG0WEenRNIIsIiIiIuJFBVlERERExIumWIiIiIiIeNEIsoiIiIiIlwCnAxytGTNm2A8//NDpGCIiIiLS9bV6YXKXG0EuLi52OoKIiIiIdGNdriCLiIiIiPiSCrKIiIiIiBefFWRjTIgxZrkxZq0xZqMx5jetHGOMMQ8bY3KMMeuMMWN9lUdEREREpC18eZFeHXCKtbbKGBMIfG2M+ae1dqnXMWcAA1tuE4EnWv4UEREREXGEz0aQbbOqlruBLbeDF10+F5jfcuxSIMYYk+yrTCIiIiIiR+LTOcjGGLcxZg1QCHxsrV120CGpQL7X/YKWxw5+nTnGmCxjTFZRUZHP8oqIiIiI+LQgW2ubrLWjgTRggjFmxEGHtLb23Pe29rPWPmWtHWetHZeQkOCDpCIiIiIizTplFQtrbTnwBTDjoKcKgHSv+2nA7s7IJCIiIiLSGl+uYpFgjIlp+TgU+CGw+aDD3gdmtaxmMQnYb63d46tMIiIiIiJH4stVLJKBF4wxbpqL+BvW2r8bY24EsNY+CSwAzgRygBrgGh/mERERERE5Ip8VZGvtOmBMK48/6fWxBW7xVQYRERERkaOlnfRERERERLyoIIuIiIiIeFFBFhGRbi09IxNjzDHf0jMynf4WRKST+fIiPREREccV5Ocxb2H2MX/+3NMGd2AaEekKNIIsIiIiIuJFBVlERERExIsKsoiIiIiIFxVkEREREREvKsgiIiIiIl5UkEVEREREvKggi4iIiIh4UUEWEREREfGigiwiIiIi4kUFWURERETEiwqyiIiIiIgXFWQRERERES8qyCIiIiIiXlSQRURERES8qCCLiIiIiHhRQRYRERER8aKCLCIiIiLiRQVZRERERMSLCrKIkJ6RiTGmXbf0jEynvw0REZEOEeB0ABFxXkF+HvMWZrfrNeaeNriD0oiIiDhLI8giIiIiIl5UkEVEREREvKggi4iIiIh4UUEWEREREfGigiwiIiIi4kUFWURERETEiwqyiIiIiIgXFWQRERERES8qyCIiIiIiXlSQRURERES8qCCLiIiIiHhRQRYRERER8aKCLCIiIiLiRQVZRERERMSLCrKIiIiIiBcVZBERERERLyrIIiIiIiJefFaQjTHpxpjPjTHfGmM2GmNua+WYacaY/caYNS23e3yVR0RERESkLQJ8+NqNwJ3W2lXGmEhgpTHmY2vtpoOO+8pae7YPc4iIiIiItJnPRpCttXustataPq4EvgVSffX1REREREQ6QqfMQTbG9AHGAMtaeXqyMWatMeafxpjhh/j8OcaYLGNMVlFRkS+jioiIiEgP5/OCbIyJAN4GbrfWVhz09Cog01o7CngEeLe117DWPmWtHWetHZeQkODTvCIiIiLSs/m0IBtjAmkuxy9ba985+HlrbYW1tqrl4wVAoDEm3peZREREREQOx5erWBjgGeBba+28QxyT1HIcxpgJLXlKfJVJRERERORIfLmKxRTgSmC9MWZNy2O/BDIArLVPAhcANxljGoEDwCXWWuvDTCIiIiIih+Wzgmyt/RowRzjmUeBRX2UQERERETla2klPRERERMSLCrKIiIiIiBcVZBERaVV6RibGmGO+pWdkOv0tiIgcE19epCciIl1YQX4e8xZmH/Pnzz1tcAemERHpPBpBFhERERHxooIsIiIiIuJFBVlERERExIsKsoiIiIiIFxVkEREREREvKsgiIiIiIl5UkEVEREREvKggi4iIiIh4UUEWaYf27jSm3cZERET8j3bSE2mH9u40BtptTERExN9oBFlERERExIsKsoiIiIiIFxVkEREREREvKsgiIiIiIl5UkEVEREREvKggi4iI+JiWhBTpWrTMm4iIiI9pSUiRrkUjyCIiIiIiXlSQRURERES8qCCLiIiIiHhRQRYRERER8aKCLCIiIiLiRQVZRERERMSLCrKIiIiIiBcVZBERERERLyrIIiIiIiJeVJBFRERERLyoIIuIiIiIeFFBFhERERHxooIsIiIiIuJFBVlERERExIsKsoj4hfSMTIwx7bqlZ2Q6/W2IiEg3EOB0ABERgIL8POYtzG7Xa8w9bXAHpRERkZ5MI8giIiIiIl5UkEVEREREvKggi4iIiIh4UUEWEREREfHis4JsjEk3xnxujPnWGLPRGHNbK8cYY8zDxpgcY8w6Y8xYX+UREREREWkLX65i0Qjcaa1dZYyJBFYaYz621m7yOuYMYGDLbSLwRMufIiIiIiKO8NkIsrV2j7V2VcvHlcC3QOpBh50LzLfNlgIxxphkX2USERERETmSTpmDbIzpA4wBlh30VCqQ73W/gO+XaBERERGRTuPzgmyMiQDeBm631lYc/HQrn2JbeY05xpgsY0xWUVGRL2KKiIiIiAA+LsjGmECay/HL1tp3WjmkAEj3up8G7D74IGvtU9bacdbacQkJCb4JKyIiIiKCb1exMMAzwLfW2nmHOOx9YFbLahaTgP3W2j2+yiQiIiIiciS+XMViCnAlsN4Ys6blsV8CGQDW2ieBBcCZQA5QA1zjwzwiIiIiIkfks4Jsrf2a1ucYex9jgVt8lUFERERE5GhpJz0RERERES8qyCIiIiIiXlSQRURERES8qCCLiIiIiHhRQRYRERER8aKCLCIiIiLiRQVZRERERMSLCrKIiIiIiBcVZBERERERL77calpERMQvVdc1sqv8AHUNHtxuQ4DLEBMaSEJkMMYcdhNYEekBVJBFRDpQekYmBfl57XqNtPQM8vN2dlAi+U5NfSOr8srZXlRFWU1Dq8eEBbnpExfOoN4RZMSGqSyL9FAqyCIiHaggP495C7Pb9RpzTxvcQWkEwASFsnR7CavyymhssmTEhTE8JZrUXqFEBgfQ6LE0NnkorKwjt6SabUVVbNpTQXJ0CJP6xTkdX0QcoIIsIiLd1tr8clKue4xlO0rpnxDOCf3jiQ0PavXYuIhghiZH0ejxsGl3BStyy/jb6l0kXvRbdpUfIDUmtJPTi4hTdJGeiIh0S29m5XPhX5aAtVw0Lo2zj0s5ZDn2FuBycVxaDFdNzuSkgfEEpw5lxgOLeGNFPtbaTkguIk5TQRYRkW7F47H85oON/OytdYzL7MWeF+4gOfroR38D3C7GZPRiz7O3Miwlip+/vY45L66ksrb1+csi0n2oIIuISLdhreXe9zfy3OJcrj6hD/OvnYDnQEW7XrNx/z5evX4Sd581lM83F/KjxxazraiqgxKLiD9SQRYRkW7BWsvv//EtLy7dyZyT+nHvzGEEuDvmx5zLZZg9tR8vzZ5IWU0DP3p0MZ9t3tchry0i/kcFWUREuoU/L8zmma93cPUJffivM4b4ZIm2Sf3i+OAnJ5IRF8Z1L2Tx6GdbNS9ZpBtSQRYRkS7vteV5PPb5Ni6dkMG9M4f5dP3i1JhQ3rrxBM4dlcKfF27h5pdXUV3X6LOvJyKdTwVZRES6tGXbS/jv9zZw0qAEfnfu8E7Z3CM0yM0DF4/m7rOG8tHGvZz/+DfkldT4/OuKSOfQOsgi8h+stRRV1ZFfeoBd5QeobWiiscnS6PEQFRpIfEQwCRHBZMaFERLodjqu9HB5JTXc+NJK0mPDeOTSMR0257gtjGmelzw4KZJbX1nNOY99zeOXjeWEAfGdlkFEfEMFWUQAqGtsYl3Bftbkl1NT3wRAbFgQ4cFuQgLduAzsP9BAfmkNHgtuY+ifEM6wlCgyYsMcTi89UU19I7Pnr8Bj4ZmrxhMdGuhIjqkDE3jvlilcPz+LK59dzt1nDeXqE/pom2qRLkwFWaSHq21oIvqES3hucS51jR4yY8MYnBRJemwYEcHf/yeiyWMpqqxj894KsvdWsqWwioTIYEIyRzmQXnoqay2/fGc9WwurmH/tBPrGhzuap098OO/cfAJ3vL6W33ywiazcMv7w45FEhjhT2kWkfVSQRXqwdQXl3PH6GmKmXkFqTCjj+8aSFBVy2M9xuwxJ0SEkRYdw4sB4tu6rYsn2Enpf8j9c9exy7pk5jP4JEZ30HUhP9cryPN5ds5s7Tx3E1IEJTscBIDIkkKeuPJ6nvtrOfR9ls3H3fh69bCwjUqOdjiYiR0kX6Yn0QE0ey4OfbOG8x7+huq6Jfa/fzcxRKUcsxwcLcLkYmhzFrEmZlH3+DKvyyjjzoa94+qvtNHm09JX4xvqC/fzm/U1MG5zALdMHOB3nP7hchhtP7s9rcyZR2+DhvMcX89AnW8Gl+foiXUmbCrIx5m1jzFnGGBVqkS6uqq6R6+dn8eAnW5l5XDIf3X4Stblr2vWaAW4XFcv/xqdzT+bEAfH8/h/fcslTS3RVv3S4/TUN3PTyShIig3ngotG4XP45z3d8n1gW3DaVM0Yk88AnW0ie9QCFFbVOxxKRNmpr4X0CuAzYaoz5gzFmiA8ziYiP7Co/wAVPfMOXW4r4/Y9G8OAlY4gO67g5kolRITx91Tjuv3AUm/dWctbDX7Fg/Z4Oe33p2Twey9w31rCvopbHLh9Lr/AgpyMdVmx4EA9fOoanrjweV3gMr67IZ+GmvVTWNjgdTUSOoE0F2Vr7ibX2cmAskAt8bIz5xhhzjTFGVyCIdAGbdlfwo8cWs6vsAM9dPZ4rJmX65OsYY/jx8Wks+OlU+iVGcPPLq7jnvQ3UNjT55OtJz/Hkom18urmQ/z57GKPTY5yO02anDU9iz9M3MTYjhi17q5i/ZCdf5xRrcxERP9bmKRPGmDjgamA2sBp4iObC/LFPkolIh1lXUM6lf11KgMvw9s0ncNIg31/UlB4bxps3TGb2iX2Zv2QnP37iG3KLq33+daV7WrKthD9/lM3MUSlc6aNf7nzJU1fN1IEJXDk5k34J4azcWcazi3ewcNNeCitrtV21iJ9p0yoWxph3gCHAi8BMa+1375m+bozJ8lU4EWm/lTtLufrZFUSHBfLq9ZNI78Q1i4MCXNx99jAm9YvjzjfXcvYjX/N/549k5qiUTssgXV9hRS0/eXU1fePD+b/zR3bp9YWjQwM5Y0Qyk/vVsya/nI27K/h2TyVx4UEMTopkcO9Iohxaz1lE/q2ty7w9ba1d4P2AMSbYWltnrR3ng1wi0gGyckuZ9exyekeF8PLsiaTEhDqS44fDerPgtqn85JVV/OTV1SzZXsI9Zw/TTnxyRI1NHm59ZTXVdY28cv3EVtfm7opiwoKYNjiRSf3iyN5XSfbeSr7ZVsI320pIjg5hcFIkAxMjCAvqHt+vSFfT1r95vwcWHPTYEpqnWIiIH1qTX87Vz60gKSqE1+ZMIvEol3DraKkxobx+w2T+vDCbv3y5ndV55Tx22Rj6ac1kOYz7FmazPLeUBy8ezaDekU7H6XAhgW5GpcUwKi2GigMN/yrLX2QXsWhLERktG/dobXGRznXYgmyMSQJSgVBjzBjgu/e1ogDtLSvipzbs2s+sZ5YRGx7EK9c7X46/E+h28V9nDGVS3zjmvrGGmY98zb3nDOfC49OcjiZ+aOHGvfzly+1cPjGDH41JdTqOz0WFBjK+Tyzj+8RSVFn3r7Kcu3EfQe4iev1gDtuLqvRLpUgnONII8uk0X5iXBszzerwS+KWPMolIO2TvreTKZ5YRGRLIK9dPJCnaP8qxt+lDEllw21Rue20NP39rHQvW78EdGe90LPEjW/ZVcucbazkuLZp7Zg5zOk6nS4gMJiEymCn949hdXsuG3fupG3MGp9z/JT8YksjPZgxmSFKU0zFFuq3DrmJhrX3BWjsduNpaO93rdo619p1OyigibZRTWMXlTy8l0O3i5dkTSevlv2/0JEeH8tr1k/j1zGEs215KynWPsya/XDvwCYUVtVzz3ApCgtw8ccXxBAf03LnqxhhSe4Vy+vAkCp64htt/OJCsnc07Vt719joKK7X5iIgvHGmKxRXW2peAPsaYuQc/b62d18qniYgDdhRXc9lflwKGV66fRJ/4cKcjHZHLZbh6Sl9OGdKb8bfM48vgMNYWlHPigHj6xYd36dUK5NhU1zVy7QsrKKup540bJpPq0IWl/shTXc7tPxzE1Sf04eFPc3hxaS5/X7eH/z1/JOdoZRiRDnWkdZC/+wkbAUS2chMRP5BfWsNlf11KQ5OHl2dPZEBi15qjmBEXRuHr/83MUckY4O/r9vDainw2763QiHIPUt/o4dZXVrFpdwWPXTaWEanRTkfySzFhQdwzcxgL7ziZQb0j+Omrq7nr7XUcqNdmPCId5bAjyNbav7T8+ZvOiSMiR2tX+QEu/etSauqbeOX6iQxO6rq/u/aLj6BPbDib9lSwKq+Mjzbu4+ucYoYmRTEgMYLEyGCNKndTtQ1N3PTSSj7PLuJ/zhvB9CGJTkfye33jw3n9hsk88PEWnvhyG6vyynjh2gkkR2vUXaS92rSTnjHmT8aYKGNMoDHmU2NMsTHmCl+HE5HD21dRy2V/Xcr+mgZevG4Cw1O6/oiby2UYkRrNlZMyOWdUCvHhwazMK+O1Ffk8uziXjzftY33Bfgoramls8jgdVzpAdV0j1z6/gi+2FPG/543k8oldb6c8pwS6Xfx8xhBeuGYCe8prueCJJdqxUqQDtHUd5NOstT83xpwHFAAXAp8DL/ksmYgcVmFLOS6urGP+dRM5Li3G6UgdyhhD3/hw+saHc6ChiR3F1WwrrGJ7cRWb9lT867jwYDdRIYGEBwcQe9rN3L8wm15hQcSGBxETFkhseNC/7ocFuTUC7Wf2VdRy00srWZNfzv0XjuL8sVry71icNCiBV66fxKxnl3HhX5bw0nVd+90kEae1tSB/t+/lmcCr1tpS/ZARcU5+aQ1XPLOMwoo6Xrh2Asdn9nI6kk+FBroZlhzFsOQorLVU1Dayr6KWsup6KmobqTjQQElVHWGDp/DY5zkcatpyUICLXmGB9AoLIq1XKEOSohiSHMmotJhO3YJbmn2+uZA731zLgfomHr1sLGeOTHY6Upc2Mi2aN26YzBXPLOOSp5bw1k0naIMRkWPU1oL8gTFmM3AAuNkYkwAcdm0ZY8yzwNlAobV2RCvPTwPeA3a0PPSOtfa3bcwj0mNt3VfJFc8so7bBw8vXT2RsRvcuxwczxhAdGkh0aOD3npt72uk0NXmoqG2gtLqespp6SqsbKKupp6y6ntLv/qyuJ6+0hs+zi/51EWBGbBhTBsQzfXACJw9O6JFLizU2edh/oIHyAw1U1jYSGJdOVV0jIQEuAtxtmpHXJuU19Tz06VaeW5zLkKRIHr1sbJe7sNRfDewdyRs3TOb8x7/h2udX8LebpxAbHuR0LJEup00F2Vp7lzHmj0CFtbbJGFMNnHuET3seeBSYf5hjvrLWnt2mpCLCyp2lzH4hiwC3i9dvmKSNAlrhchliwoKICTtyKahtaCKnsIqs3FK+zinhg7W7eXV5HtGhgZw5MpkLjk/r1qPz+w808NnmfazcWcaqneVs3lvxH6PvKbOf4Jmvm8cwwoPdxIYF0Ss8iISI5k0s4sKDjqo4F1fV8fRXO3hxSS7V9U3MmpzJL88cSkhgz/tlxJcy48J5atY4Lv3rUm54MYuXZk/skb/wibRHW0eQAYbSvB6y9+ccsvxaaxcZY/ocazAR+U9vZuXzq79tICUmhBeunUBmnP+vc+zvQgLdjEiNZkRqNFdP6UtDk4evc4p5b/Uu3l29i1eX5zEmI4bZJ/bj9OG9O3QU1Skej2XR1iLeWlnAwk37qG/0EBEcwOj0GG6a1p/eUSFEhwYSERzAOedfwMW/mMeB+ibKDzSPvG/eU8m6pv0AuAz0Cg8iMTKYhIhgIkMCCQ92ExroxmMhKHkQn28uZOXOMpZsL2FtfjlN1nL2cSncOn2A5sj60PGZvbj/wlH85NXV/OKtdTxw8WjNvxc5Cm0qyMaYF4H+wBrgu4UWLYcfHW6LycaYtcBu4P9Zaze28/VEup0mj+WPH27mqUXbmTIgjscuG9um0VE5eoFuF9MHJzJ9cCLVdY28vaqAZ77ewS2vrKJPXBh3nDqImcel4HJ1vaJhreWjjft48JMtbN5bSUxYIJeOT+e8sWmMTI3G3cr3VLP5a0YetBaxtZb9BxooqqyjqKqOwso6dpbU8O2eyu99fvKseVzz/ArcLSuTzJ7ajwvHpWlebCeZOSqF3OJq7v94C+P6xHLFJK0OItJWbR1BHgcMs9Z25Ir9q4BMa22VMeZM4F1gYGsHGmPmAHMAMjIyOjCCiH8rKKth7utrWZ5byqzJmfz32cMI7AajmF1BeHAAsyb34fKJmXy8aS8PfrKV215bw5NfbueuM4Zw8qAEpyO22ZJtJfz+H5vYuLuCfvHhPHDxKM4cmXxMb7sb8+8pLAN7/3sEuKa+keq6JqrrGqlpaMJtDPN/fSNffvoRQ5IiiQz5/pxx8b1bpg9geW4pv/v7Jib2jf2PcyYih9bWn7QbgKSO/MLW2gprbVXLxwuAQGNM/CGOfcpaO85aOy4hoev8UBJpj/fX7uaMh75i054K5l00it+eO0Ll2AFul2HGiGQW/HQqD10ymuq6Rq56djlz5mdRUFbjdLzDKqys5fbXVnPpX5dSXtPA/ReOYuEdJ3HemLQOn5MaFhRAQmQwfeLDGZYcxeCkSA5sz2J8n1iVYwe5XIb7LxxFeHAAP31tDbUN2m1PpC3aOoIcD2wyxiwH6r570Fp7zrF+YWNMErDPWmuNMRNoLuslx/p6It1FQVkNv/v7Jj7auI+xGTE8ePEYMuK0BJnTXC7DuaNTOWNEMk9/vZ1HPs3hh/O+5Kc/GMjsE/sRFOA/v7xYa3lleR5/WLCZukYPPz1lADdNG0BokC7U6okSo0K474LjuO6FLP70YTb3zBzmdCQRv9fWgvzro31hY8yrwDQg3hhTANxLy3rK1tongQuAm4wxjTQvH3dJB0/hEOlSahuaeObrHTzy2VYMhp/PGMycqf26xYVh3UlQgIubpw3g3NGp/PaDjfzpw2zeXlnA784dwQkDWn0TrFPlldRw1zvr+GZbCVMGxPH7H42kb3zXvaAzPSOTgvw8p2N0eT8Y2purJmfy7OIdnD68NxP7xTkdScSvtXWZty+NMZnAQGvtJ8aYMOCwQxHW2kuP8PyjNC8DJ9KjmYBgnv16B09+uY3CyjrOGJHE3WcPIzUm1OlochipMaH85cpxfJ5dyK/f38hlTy9j5qgU3OHOLAvn8VheWJLLnz7MJsBl+MP5I7l4fHqXX7mgID+PeQuz2/Uac08b3EFpurZfnDGETzcX8qt3N7Dgp1P96l0PEX/Tpr8dxpjrgbeAv7Q8lErzRXUicowamjys3FlG6g1P89u/b6JfQjivXD+RJ644XuW4C5k+OJGPbj+J2384kI827iXl+idZm1+OpxPfENteVMVFf1nCbz7YxKR+sXx0x0lcMiGjy5djv2FcGGPadfMHYUEB/O7cEeQUVvHUom1OxxHxa22dYnELMAFYBmCt3WqMSfRZKpFurL7Rw9qCclbnlXOgoYn6oh18cOeZTOgb63Q0OUYhgW5u/+EgfjQ6lYm3zOOL4HA27angpEEJPv1lp66xib8u2s4jn+UQEuhm3kWjOG9Mqt8Usm7DerrNKPb0IYmcNTKZhz/L4ezjUujThaffiPhSW99fqbPW1n93p2WzEM0XFjkKdY1NLN9RynOLd/DNthISo4K5aFwahW/co3LcTfSJD6fwjXuYMTyJmvom3lpZwD/X76HiQEOHf60vsgs5/YFF/HnhFn4wNJGP7ziJ88emqRzLEd0zcxjBbhd3v7sBXfoj0rq2jiB/aYz5JRBqjDkVuBn4wHexRLqP+kYPa/LLWZlXRn2jh77x4UzoG0tSVIjT0cRHBidF0i8hnKydZazcWca2omqGpkQyLjOW6ND2LXm2IreUhz/dyldbi+kXH878aydwUhdak1mc1zsqhJ/NGMw9723ko417mTEi2elIIn6nrQX5LuA6YD1wA7AAeNpXoUS6gyaPZcOu/SzbUcqBhib6xYczsV8siZEqxj1BoNvF5H5xjEiJYkVuGZt2V7BxdwWDEiMZmhxJeq+wNu/IV9vQxBfZRby4NJfFOSXERwTxqzOHMuuEzA5fz1h6hssmZPDS0p383z83c8qQ3rpgT+QgbV3FwmOMeRd411pb5NtIIl3f7vIDfLq5kNLqelJjQpkyII7kaF141xNFhgRyypBEJvSJZWVec1HO3ldJWJCb/gkRJEWFkBAZTK/wQNzfXdQVGMy6gnKy91aybEcpH23YS2VdIwmRwdx91lAun5ipNY2lXQLcLn511jCuenY585fkMntqP6cjifiVwxZk0zyZ7V7gVsC0PNQEPGKt/W0n5BPpUuoam/gmp4R1u/YTERzAzOOS6RsfrnmhQkRIACcPSmBK/zh2lFSTvbeSzXsrWL9r//eOzZj7Nuc8uhiAyOAATh+RxDmjUjihf5zWxZYOc/KgBE4alMDDn27lx2PT6BUe5HQkEb9xpBHk24EpwHhr7Q4AY0w/4AljzB3W2gd8nE+kyyiuquMf6/aw/0ADo9NjmNwvTm9byvcEuF0MTIxkYGIkHmvZX9NAYWUd+w804LEWa2Hh/Id47akHGNQ7ksy4cNxtnIohcrR+deZQznhoEQ9/tpV7Zw53Oo6I3zhSQZ4FnGqtLf7uAWvtdmPMFcBCQAVZBNi8t4JPvy0kKMDFj8emkdqr86ZTaKexrstlDL3Cg743cvfmkteZMeI1h1JJTzI4KZJLJmTw4pKdXDW5j5Z9E2lxpIIc6F2Ov2OtLTLGtO9SbJFuYnFOMVk7y0iJCeHMEcmEB7f12teOoZ3GRKQ9bv/hQN5ZVcCDn2zhwUvGOB1HxC8c6f3f+mN8TqTb83gssafeSNbOMkakRHH+mLROL8ciIu2VGBnC1Sf05b21u8neW+l0HBG/cKSCPMoYU9HKrRIY2RkBRfxRY5OH//fmWiLHns3xGb04ZUii5om2czteEXHOjSf3IyIogPvb+W6USHdx2OEua63WERI5iLWWn7+1jndW76J80YtM+fVvVPCg3dvxapqHiHNiwoK4/qR+zPt4C2vyyxmdHuN0JBFH6RJ7kaP054XZvLN6F3NPHcT+Ja+rHItIt3DtiX2JDQ/SKLIIKsgiR+XlZTt57PNtXDohnZ+cMsDpOCIiHSYiOICbTu7PV1uLWba9xOk4Io5SQRZpo883F/Lf725g+uAEfnfuCI0ci0i3c8WkTOIjgnno061ORxFxlAqySBvkl9Zw++trGJIUxaOXjdVuZiLSLYUGubnx5H58s62EFbmlTscRcYx+yoscQX2jh1tfXY3HY3niirFayk1EurXLJ2YSHxHEQ59oFFl6LhVkkSP4v39+y9r8cu678Dgy47TLlIh0b6FBbuac1I+vc4pZuVOjyNIzqSCLHMaHG/by3OJcrpnShxkjkp2OIyLSKa6YlElceBAPahRZeigVZJFDKK6q45d/W8/I1Gj+64yhTscREek0YUEBXH9SP77aWsyqvDKn44h0OhVkkUO4570NVNU2Mu+iUQQF6K+KiPQsV07KJDZcc5GlZ9JPfZFW/H3dbhas38vtpw5kYO9Ip+OIiHS68OAAZk/ty5dbiliTX+50HJFOpYIscpDiqjrueW8jo9KimTO1n9NxREQcM2tyH2LCAnnoky1ORxHpVCrIIgf57QebqKpt5M8XjtJ6xyLSo0UEB3D91H58nl3EWo0iSw+in/4iXhbnFPP+2t3cNK2/plaIiH8xLowx7bqlZ2Qe9ZedNTmT6NBAHtbuetKDaMcDkRb1jR7ueW8DmXFh3DStv9NxRET+k/Uwb2F2u15i7mmDj/pzIkMCmX1iX+7/eAvrC/YzMi26XRlEugKNIIu0ePrr7WwrqubX5wwnJNDtdBwREb9x1ZQ+RIUEcPZdjzkyii3S2TSCLALsKj/AI5/mcPrw3kwfnOh0HJF2S8/IpCA/z+kY0k1EhQRy3Yn9eKC2kbveXktiZMgxv9axjGKLdDYVZBHgf/6xCYB7Zg53OIlIxyjIz3Pk7Xjpvq6e0oc//30Vy3eUcvZxKU7HEfEpTbGQHm9FbikL1u/lxpP7kxoT6nQcERG/FB0aSGXWe2wrqqaoss7pOCI+pYIsPZrHY/nd3zeRHB3CnJO05nGX59BV/iI9RUXW+wQFuFi2o8TpKCI+pSkW0qO9u2YX6wr288DFowgN0oV5XZ5DV/mL9BS2rprR6TEs31FKUWUdCZHBTkcS8QmNIEuPdaC+iT99mM1xadGcOyrV6TgiIl3CmPQYgtwulueWOh1FxGdUkKXH+utX29lbUcvdZw3D5TJOxxH5tw6YKiLiKyGBbkanx5BTWEVxleYiS/ekKRbSI5VU1fHUou2cPrw3E/rGOh1H/ElLOXWUpoqInxuTEcOa/HKW7yjlzJHJTscR6XAqyNIjPf7FNmrqG/nZ6SoRcpB2llMVU+kJQgLdjEqPZkVuGSVVdcRFaC6ydC+aYiE9zq7yA7y4dCc/HpvGgMRIp+OIiHSedk7f8TYmoxeBbqO5yNItaQRZepyHPtkCFm4/dZDTUUREOlcHvkMSGuhmVFoMWTvLmNBHo8jSvWgEWXqUnMJK3lpZwBWTMrUpiIhIO43N7EWQ28WS7VoXWboXFWTpUe5fuIXQQDe3TO/vdBQRkS4vNNDN8Zm92FZUzZ79B5yOI9JhfFaQjTHPGmMKjTEbDvG8McY8bIzJMcasM8aM9VUWEYC1+eX8c8NeZk/tp7cCRUQ6yJiMGMKC3CzOKcFa63QckQ7hyxHk54EZh3n+DGBgy20O8IQPs4hw30fZxIYHMXtqX6ejiIh0G4FuFxP6xrKr/AC5JTVOxxHpED4ryNbaRcDhLm09F5hvmy0FYowxWkxRfGJxTjFf5xRz87T+RIYEOh1HRKRbGZESTXRoIN9sK9YosnQLTs5BTgXyve4XtDwm0qGstfzpw82kRIdwxaRMp+OIiHQ7bpdhcr84iqvq+XZPpdNxRNrNyYLc2lZVrf7aaYyZY4zJMsZkFRUV+TiWdDcfbdzL2oL93H7qIEIC3U7HERHplgb1jqB3VDDfbCumvtHjdByRdnGyIBcA6V7304DdrR1orX3KWjvOWjsuISGhU8JJ99DY5OG+j7LpnxDO+WP0BoWIiK8YYzh5UALV9U1k7dTmIdK1OVmQ3wdmtaxmMQnYb63d42Ae6YbeWb2LbUXV/Oz0wQS4taqhiIgvJUeHMjgpklV55ew/0OB0HJFj5stl3l4FlgCDjTEFxpjrjDE3GmNubDlkAbAdyAH+CtzsqyzSM9U2NPHgx1sYlRbN6cOTnI4jItIjTOkfhwG+zil2OorIMfPZVtPW2kuP8LwFbvHV1xd5eVkeu/fXct+FozCmtSnvIiLS0SJDAhnXpxdLt5eSX1pDemyY05FEjprec5Zuqaqukcc+z2HKgDimDIh3Oo6ISI9yfEYvokMD+Ty7kEaPLtiTrkcFWbqlp7/aTml1PT8/fYjTUUREepwAt4vpgxMoq2kgK7fM6TgiR00FWbqdoso6/rpoOzOGJzEqPcbpOCIiPVJmXDiDekeQlVtGWXW903FEjooKsnQ7j362ldpGDz+bMfiwx6VnZGKMaddNREQO7aSBCQS4DZ9tLtQOe9Kl+OwiPREn7Cyp5uVleVw8Pp3+CRGHPbYgP495C7Pb9fXmnnb4Ei4i0pOFBwcwZUA8n20uZOPuCkakRjsdSaRNNIIs3cqfF24h0O3i9h8MdDqKiIgAI1KiSOsVyqKtRVobWboMFWTpNtYVlPPB2t3MntqXxKgQp+OIiAjNO+ydOqw3BsPCTXvBqHqI/9P/pdItWGv5wz830ysskDkn9XM6joiIeIkKCWTa4AR2l9cSNf5cp+OIHJEKsnQLX20t5pttJfzklIFEhgQ6HUdERA4yJCmS/gnhxEydxbd7KpyOI3JYKsjS5Xk8zaPH6bGhXD4pw+k4IiLSCmMMpwxJpKm2klteWUVVXaPTkUQOSQVZurwP1u1m054K/t9pgwkOcDsdR0REDiEsKIDi9+8jt7iaX/1tvZZ+E7+lgixdWl1jE/d9lM3wlChmHpfidBwRETmCuvz1zD11EO+t2c0ry/OcjiPSKhVkcUR7N+lIz8gE4OWleRSUHeCuM4bgcmnjDhGRruDmaQM4aVACv/lgE+sL9jsdR+R7tFGIOKK9m3TMPW0w5TX1PPzZVk4cEM/UgQkdmE5ERHzJ5TI8cNEoZj7yNXNezOK9W6eQGKnlOcV/aARZuqwHP9lKxYEG7j57qNNRRETkKMVFBPPUrHGU1zRw44srqWtscjqSyL+oIEuXFBCXxotLd3LphAyGJEU5HUdERI7BiNRo7r9oFKvyyvnlOxt00Z74DRVk6ZJiT5lNWJCbuacOcjqKiIi0w5kjk7ntBwN5e1UBj3+xzek4IoDmIEsXtKO4mtB+47jtBwOJiwh2Oo6IiLTTbT8YSG5JNfd9lE1iZDAXjkt3OpL0cCrI0qU0ejws2lpEQ+kuZk0+w+k4IiLSAVwuw30XjKK0up673llPXEQQpwzp7XQs6cE0xUK6lNV55ZTXNFD6yV8ICtD/viIi3UVQgIsnrjieYclR3PzyKlbllTkdSXowNQzpMipqG1i+o5T+CeHU7ljldBwREelgEcEBPHfNeHpHhXDt8yvIKaxyOpL0UCrI0mUs2lIEwEmDEsC42rXRiDHaVERExB/FRwQz/9oJBLgMVz27nL37a52OJD2Q5iBLl5BbXM22ompO6B9HVEggWE+7NhqB5s1GRETE/2TGhfP8NRO4+C9LuOrZ5bxx42SiQwOdjiU9iEaQxe81NHn4YksRvcICGZvRy+k4IiLSCUakRvOXK8exvbiK655fwYF6bSQinUcFWfzesh2l7D/QwClDEnG7NDVCRKSnOHFgPA9ePIZVeWXc+NJK6hs9TkeSHkIFWfxaYUUtq/LKGJ4SRVqvMKfjiIhIJzvruGT+97yRfLmliDveWEOTR7vtie9pDrL4LY/H8snmQkID3Zw4IN7pOL7TcsGhiEiPcIz/5kVNOJ9/cC2vzX+OsE3vk5+30wfhRJqpIIvfWp1fTlFlHWeOSCIk0O10HN9p5wWHuthQRLqUdvybtzinmCxmsL9Wy7+Jb2mKhfilsup6lmwvoV98OAMSI5yOIyIifuCE/nGMTI0metIFPPHFNqfjSDemgix+x+OxLNy0j0CX4ZQhiZp+ICIiABhjmD44gepNX/DHDzfz8jJNsxDfUEEWv7Myr4y9FbVMG5xIeLBmAYmIyL8ZYyj+xwOcMiSRu9/dwPtrdzsdSbohFWTxK8VVdSzdXsLAxAgG9dbUChERaYWniccuG8v4zFjmvr6GzzcXOp1IuhkVZPEbjR4PCzfuIzjAzfTBmlohIiKHFhrk5umrxzEkOZIbX1rJitxSpyNJN6KCLH5jybYSiqrq+OHQREKDuvGqFSIi0iGiQgJ54ZoJpMaEMvuFLHIKK52OJN2ECrL4hZ0l1azKK2dkajT9EjS1QkRE2iYuIpgXrp1AoNvFVc+uYF9FrdORpBtQQRbH1dQ3snDTPmLDgpg6sBtvCCLS07RsCNGem0hbpMeG8fw14ymvqeeqZ5dTWdvgdCTp4rREgDjKWssn3xZS1+DhR6NTCXTrdzaRbqOdm+CANsKRthuRGs3jVxzPdc+v4MaXVvLc1RMICtDPFDk2+j9HHLU6v5wdxdVMGRBHQmSw03FERKQLO3lQAn/48XEszinhF2+vw1rrdCTpojSCLI7Zs/8Ai3OK6Z8Qzuj0GKfjiIhIN3DB8Wns3X+APy/cQlJ0CL+YMcTpSNIFqSCLI1whESxYv5eI4ABOHdpbcw1FRKTD3DJ9ALv31/LEF9tIjg5h1uQ+TkeSLkYFWTqdx2OJO2suNfWNXDQuneBALekmIiIdxxjDb88ZTmFFLfe+v5HEyBBmjEhyOpZ0IZqDLJ3u4c+2EjZgAicNTKB3VIjTcUREpBsKcLt45NKxjEqL4bbXVrNypzYSkbbzaUE2xswwxmQbY3KMMXe18vw0Y8x+Y8yalts9vswjzvt40z4e/GQrVes/4bi0aKfjiIhINxYa5ObZq8eTEhPKdS9kkVNY5XQk6SJ8VpCNMW7gMeAMYBhwqTFmWCuHfmWtHd1y+62v8ojzthVVMff1NYxMjaZ04eOadywiIj4XGx7EC9dMIMBluPq55RRW1pKekdnuNbrTMzKd/tbEh3w5B3kCkGOt3Q5gjHkNOBfY5MOvKX6qsraBOfOzCAxw8eSVx5P203qnI4mISA+RERfGs1eP55KnlnLNcyvYta9Ia3TLYflyikUqkO91v6DlsYNNNsasNcb80xgzvLUXMsbMMcZkGWOyioqKfJFVfMjjsdz5xlpyS2p49LIxpMaEOh1JRER6mOPSYnjs8rFs3ltJwrl30eTRGslyaL4syK29f37w/42rgExr7SjgEeDd1l7IWvuUtXactXZcQkJCx6YUn3v8ixwWbtrHL88cygn9tZW0iIg4Y/rgRP73vBGE9jueTzfv00Yicki+LMgFQLrX/TRgt/cB1toKa21Vy8cLgEBjjBpUN/L55kLu/3gLPxqdwrVT+jgdR0REeriLx2dQ/vUrfLunkqXbtbKFtM6XBXkFMNAY09cYEwRcArzvfYAxJsm0XKlljJnQkqfEh5mkE+UUVvLT11YzNCmK/zv/OF2UJyIiHcO42nWB3f7FrzA8JYrluaVs2LXf6e9G/JDPLtKz1jYaY24FPgLcwLPW2o3GmBtbnn8SuAC4yRjTCBwALrF6v6NbKK2u59rnswgOcPHUrOMJDdJmICIi0kGsp10X2c09bTDTBydSVdfIZ9mFhAW76Rcf0YEBpavz6U56LdMmFhz02JNeHz8KPOrLDNL56hqbuOHFLPZW1PLanEmk9QpzOpKIiMh/cLsMZ45I5u1VBSxYv5dzR6WQHqufV9JMO+lJh7LW8l/vrGdFbhn3XziKsRm9nI4kIiLSqqAAFz8ak0pMaCAfrNvN3v21TkcSP6GCLB3q8S+28c6qXdzxw0HMHJXidBwREZHDCg10c96YVMKCAnh3zS6KKuucjiR+QAVZOsyC9Xu476Nszh2dwk9/MMDpOCIiIm0SHhzA+WNSCXS7eHfNLspqtJlVT6eCLB1ibX45c99Yw9iMGP74Y61YISIiXUtUaCDnjUnFWnhn1S4qahucjiQOUkGWdssvrWH2/CziI4J5atY4QgK1YoWIiHQ9seFBnDcmlfomD++s2kV1XaPTkcQhKsjSLmXV9Vz13HLqGpp49urxxEcEOx1JRETkmCVEBvOj0SnU1Dfy9qoCqlSSeyQVZDlmB+qbuO6FFRSUHeDpq8YzqHek05FERETaLTk6lHNHpVJV18hbKws03aIHUkGWY9LY5OEnr65mdX45D18ymgl9Y52OJCIi0mFSe4Vy3phUDtQ38dbKAvYfUEnuSVSQ5ahZa7nn/Y188u0+fnPOcGaMSHY6koiISIdLjg7l/LGp1Dd6eCMrX0vA9SAqyHLUHvksh1eW5XHztP7MmtzH6TgiIiI+0zsqhAuPT8NlDG+tLCCvtMbpSNIJVJDlqLyxIp95H2/h/LGp/Oz0wU7HERER8bm4iGAuGpdGZEgA763ZxeY9FU5HEh9TQZY2W7B+D//1t/WcNChBax2LiEiPEhkSyIXHp5ESHcpHm/YRc/JVNHms07HER1SQpU0+2bSPn766mjHpMTxx+VgC3fpfR0REepbgQDc/GpPKyNRooiddyPXzs7TCRTelliNH9OWWIm5+eRXDU6N57prxDBnYH2NMu24iIiJdkdtlOGVIIiUfPcaiLUWc++hiNuza73Qs6WABTgcQ//bNtmLmzM9iQGIE86+ZQGRIIAX5ecxbmN2u1517muYvi4hI11W15p98+s5L/PTV1Zz/+Df86qyhzJqcqUGgbkIjyHJIK3JLue75LDLjwnhp9kSiwwKdjiQiIuI3JvSNZcFtUzlxYDz3vr+R6+evpLCi1ulY0gFUkKVVa/LLuea5FSRHh/DS7InEhgc5HUlERMTvxIYH8fSscdx91lC+2lrEqQ8s4t3Vu7BWF/B1ZSrI8j2r8sqY9cwyYsODeOX6SSRGhjgdSURExG+5XIbZU/ux4Lap9E8I5/bX13DdC1nsLKl2OpocIxVk+Q/fbCvmiqeX0Ss8iFeun0hStMqxiIhIW/RPiODNG0/g7rOGsmx7Cac+sIh5C7M5UN/kdDQ5SirI8i+fbd7HNc+tIDUmlDdvmExarzCnI4mIiHQp7pbR5E/vnMYZI5J4+LMcfnD/F7yZla91k7sQFWQB4O2VBcyZv5KBvSN4/YbJJEZp5FhERORYJUWH8NAlY3h9ziQSIoP52VvrOOOhRXy0cS8eFWW/p4Lcw1lrefSzrdz55lom9I3llesn6YI8ERGRDjKxXxzv3jKFxy8fS2OT5YYXV3L6g4t4e2UBDU0ep+PJIagg92ANTR5++bcN/HnhFn40OoXnr5lAVIiWchMREelIxhjOHJnMwjtO4qFLRuN2Ge58cy3T7vuC5xbv0BxlP6SNQnqo0up6bn55JUu3l3LztP787PTBWtxcRETEhwLcLs4dnco5o1L4IruIx7/I4TcfbOKRz3K4clIml0/M0BRHP6GC3ANt2l3BnBezKKys44GLR3HemDSnI4mIiPQYxhimD0lk+pBEVuSW8uQX23jo06089nkOpw9P4srJmUzsG6uBKwepIPcg1lreXFnAve9tJDo0kDdvmMyo9BinY4mIiPRY4/vEMv7qWHKLq3l52U7eyCrgH+v3MKh3BFdOyuS8sWlEBKuudTb9F+8hquoa+dXf1vPemt1M7hfHQ5eO1gYgIiIifqJPfDi/OmsYc08dzAdrdzN/aS7//d5G/vDPzcwclcIlEzIYlRatUeVOooLcA2TllvL/3lxLXmkNd546iJunD8Dt0l8wERERJ6VnZFKQn3fI54OSBxE5+gxeqZrKayvyqS/cQdXaj6ja+Dm2rpq09Azy83Z2YuKeQwW5G6upb+RPH2bzwpJcUmNCef2GyYzvE+t0LBEREQEK8vOYtzD7iMfVNTaRvbeSjZHBFCb2JeH0mxiYGMEX825p94iySnbrVJC7IWstn35byG/+vpH80gNcfUIffnb6YMI1h0lERKTLCQ5wc1xaDMelxVBYUcuG3RVk760k6bI/0CsskBEp0QxJjiQs6Oh/zs89bbAPEnd9akzdTE5hJb/5YBNfbS1mQGIEb9wwmQl9/z1qfKS3c0RERMR/JUaFcEpUCFMHxvPbn/+E5Evv4qucYhZvK6Z/QgQjUqNJ7xWqucrtpILcxRyq4LqjEok+4WIiRv4QW3+A8q9f4dPV/2Dind9ffLwtb+ccjn7bFBERcVag20X1+k+46P7HKK6qY+PuCr7dU8HWwiqiQwMZlhLF8OQovXt8jPRfrYs5eL5SWU09q3aWsWlPBcYYRqREMaFvLGFn3Qfc973PV7kVERHpXuIjgjl5UAJT+seRU1TFxl0VLNlWwtJtJWTGhTEsOYq+CeEEuLSBclupIHdB1lrySmtYk19ObkkNbmMYkRLNuD69iNRW0SIiIr5nXH43jSHA7WJIUhRDkqIoq6ln0+4KNu+tZMGGvYQEuBicFMnQ5CgSI4P9Lru/UUHuQrYXVRE99Qqe+yaXytpGQgPdTOwby8jUaL2FIiIi0pmsx6+nLPYKC2LKgHgm948jv7SGTXsq2LC7grUF+4kLD2JQ70gG9Y7w2dfv6tSq/JjHY9mwez8fb9rHx5v2sXlvJdGTLiQ2LIgp/ePpn6i3S0REROTQXMaQGRdOZlw4dQ1NbNlXxbd7K1iyvYQl20tImjWPvy7azlnHJZMSE+p0XL+hguxH6hs9ZO+tZFVeWfPcoR0llNc04DIwrk8sd581lBvPHM/t7yxxOqqIiIh0McGBbkamRTMyLZrK2ga27qvikz1b+J8F3/I/C77l+Mxe/HBob04Zksig3hE9ehqGCnIbdcTyaN8txt3ksRSU1bCtqIqcwiq2FVaTva+STXsqqG/0AJAaE8qpQ3tzwoA4Th6USGx4EADXV5W2+3sRERGRni0yJJCxmb146fq57Ciq4u/rdvPPDXv544eb+eOHm0mNCeUHQxM5ZUgik/rFERLodjpyp1JBbqO27nZjraW+0UNVXeP3bsu//IQZDy5iR3E1dS1FGCA+Ioj+CRFcfUIfRqXFMCo9mrReYb78dkREREQA6BMfzq2nDOTWUwayd38tn2cX8um3hbyRlc/8JTsJCXQxNqMXE/rGMqFvLGPSexEa1L0LswryUfB4LDX1Td8rvtXffVzb/Gejx37vc0MD3QRExpMaE8rUgfEMSIygf0IEAxIjiAkLcuC7EREREflPSdEhXDohg0snZFDb0MSS7SV8mV3EitxSHvp0K9ZCoNswMjWa8X1jGZESzfCUKPrEheNydZ8pGSrIbfD84h2k3vwCj36ew8HV12UgIjiA8OAAEiKD6Rsf/q/7ESEBLR+7CXC5mHvaGTzz3PfLs4iIiIgjjmK5OhMcTnDqUELShvNN+nBW7hiIcTcvLxsW5GZIUiTDUqIYmhxF37hwMuPDSY4K6ZLF2acF2RgzA3gIcANPW2v/cNDzpuX5M4Ea4Gpr7SpfZjoWKTGh1O5YyQnnXNJceEPcRAQ3l9/QQHfbJ7H74ZqJIiIi0oO1Y7m6Ro+HsuoGHr73Dn72vw+xcXcF763ezUtL/33NVlCAi/ReofSJCyc9Noyk6BB6RwWTGBlCYmTzn1GhAX7Xj3xWkI0xbuAx4FSgAFhhjHnfWrvJ67AzgIEtt4nAEy1/+pXThidR8s+HmXzHLe17IT9fM1FERESkrQJcLhIig6ne8Cn3zhwONF+Ltav8AHklNeSW1LCzpJrckmp2ltSwZHsJNfVN33ud4AAXf7nyeKYNTuzsb+GQfDmCPAHIsdZuBzDGvAacC3gX5HOB+dZaCyw1xsQYY5KttXt8mEtEREREOsrRTNMICsUdEYs7PBZ3RC/cEXG4I3px5d/uIX/jCh8HbTvT3E198MLGXADMsNbObrl/JTDRWnur1zF/B/5grf265f6nwC+stVkHvdYcYE7L3cFA+4Zhm8UDxR3wOuL/dK57Bp3nnkPnuufQue45nDrXxdbaGQc/6MsR5NZ+lTi4jbflGKy1TwFPdUSof31hY7KsteM68jXFP+lc9ww6zz2HznXPoXPdc/jbufblPsUFQLrX/TRg9zEcIyIiIiLSaXxZkFcAA40xfY0xQcAlwPsHHfM+MMs0mwTs1/xjEREREXGSz6ZYWGsbjTG3Ah/RvMzbs9bajcaYG1uefxJYQPMSbzk0L/N2ja/ytKJDp2yIX9O57hl0nnsOneueQ+e65/Crc+2zi/RERERERLoiX06xEBERERHpclSQRURERES8dPuCbIyZYYzJNsbkGGPuauV5Y4x5uOX5dcaYsU7klPZpw3m+vOX8rjPGfGOMGeVETmm/I51rr+PGG2OaWtZkly6oLefaGDPNGLPGGLPRGPNlZ2eU9mvDv9/RxpgPjDFrW85zZ16vJB3IGPOsMabQGLPhEM/7TSfr1gXZa7vrM4BhwKXGmGEHHea93fUcmre7li6kjed5B3CytfY44Hf42cUA0jZtPNffHfdHmi8Sli6oLefaGBMDPA6cY60dDlzY2Tmlfdr4d/oWYJO1dhQwDbi/ZXUs6XqeB763KYcXv+lk3bog47XdtbW2Hvhuu2tv/9ru2lq7FIgxxiR3dlBplyOeZ2vtN9baspa7S2lec1u6nrb8nQb4CfA2UNiZ4aRDteVcXwa8Y63NA7DW6nx3PW05zxaINM17GUcApUBj58aUjmCtXUTz+TsUv+lk3b0gpwL5XvcLWh472mPEvx3tObwO+KdPE4mvHPFcG2NSgfOAJzsxl3S8tvy9HgT0MsZ8YYxZaYyZ1WnppKO05Tw/CgyleSOx9cBt1lpP58STTuY3ncyXW037gw7b7lr8WpvPoTFmOs0F+USfJhJfacu5fhD4hbW2qXnASbqotpzrAOB44AdAKLDEGLPUWrvF1+Gkw7TlPJ8OrAFOAfoDHxtjvrLWVvg4m3Q+v+lk3b0ga7vrnqFN59AYcxzwNHCGtbakk7JJx2rLuR4HvNZSjuOBM40xjdbadzsloXSUtv77XWytrQaqjTGLgFGACnLX0ZbzfA3wB9u8cUOOMWYHMARY3jkRpRP5TSfr7lMstN11z3DE82yMyQDeAa7U6FKXdsRzba3ta63tY63tA7wF3Kxy3CW15d/v94CpxpgAY0wYMBH4tpNzSvu05Tzn0fwuAcaY3sBgYHunppTO4jedrFuPIHeB7a6lA7TxPN8DxAGPt4wsNlprxzmVWY5NG8+1dANtOdfW2m+NMR8C6wAP8LS1ttXlo8Q/tfHv9O+A540x62l+C/4X1tpix0LLMTPGvErzSiTxxpgC4F4gEPyvk2mraRERERERL919ioWIiIiIyFFRQRYRERER8aKCLCIiIiLiRQVZRERERMSLCrKIiIiIiBcVZBERERERLyrIIiIiIiJe/j+7mZXUF6IyiAAAAABJRU5ErkJggg==\n",
      "text/plain": [
       "<Figure size 720x360 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 720x360 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "num_attr = ['hum', 'atemp', 'temp', 'cnt']\n",
    "\n",
    "for i in num_attr:\n",
    "    sns.displot(df[i], kde=True, stat='density', aspect=2)\n",
    "    plt.title(i, size=15, weight='bold')\n",
    "    plt.xlabel(None)\n",
    "    plt.show()\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### **Modeling**\n",
    "---"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**Custom Transformer to Extracting `dteday` Feature**\n",
    "\n",
    "Untuk memperkaya data dan mempermudah dalam melakukan interpretasi maka fitur `dteday` akan kita ekstrak menjadi bagiannya sendiri, fitur baru hasil ekstraksi beserta keterangannya bisa dilihat tabel dibawah.\n",
    "\n",
    "|Atribute|Data Type|Description|\n",
    "|---|---|---|\n",
    "|year|Integer|Year (2011 to 2012)|\n",
    "|month|Integer|Month (1 to 12)|\n",
    "|day|Integer|Day (1 to 31)|\n",
    "|weekday|Integer|Day of the week (0: Sunday, 1: Monday, 2: Tuesday, 3: Wednesday, 4: Thursday, 5: Friday, 6: Saturday)|\n",
    "|workingday|Integer|If the day is neither weekend (saturday & sunday on weekday column) nor holiday is 1, otherwise is 0|"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [],
   "source": [
    "class DateTransformer(BaseEstimator, TransformerMixin):  \n",
    "    \n",
    "    def fit(self, X, y=None):\n",
    "        return self\n",
    "\n",
    "    def transform(self, X, y=None):\n",
    "        X['year'] = X['dteday'].dt.year\n",
    "        X['month'] = X['dteday'].dt.month\n",
    "        X['day'] = X['dteday'].dt.day\n",
    "        X['weekday'] = X['dteday'].dt.day_of_week\n",
    "        X['workingday'] = np.where((X['holiday'] == 1) | (\n",
    "            X['weekday'] == 5) | (X['weekday'] == 6), 0, 1)\n",
    "        X = X.drop('dteday', axis=1)\n",
    "        X = X.astype(str)\n",
    "        return X"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**Data Spliting**\n",
    "\n",
    "Pada tahap ini kita melakukan pembagian data, disini kita tentukan fitur target yaitu adalah `cnt` dan sisanya akan masuk pada pemodelan untuk menemukan pola, selanjutnya akan ada pemisahan antara data untuk di latih dan data untuk melakukan tes/validasi"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {},
   "outputs": [],
   "source": [
    "X = df.drop(['cnt'], axis=1)\n",
    "y = df['cnt']\n",
    "\n",
    "X_train, X_test, y_train, y_test = train_test_split(\n",
    "    X,\n",
    "    y,\n",
    "    test_size=.2,\n",
    "    random_state=42)\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**Encoding** \n",
    "\n",
    "kita perlu melakukan encoding pada fitur kategorik karena komputer tidak dapat memproses data berbentuk kategorik, walaupun nilainya sudah berupa angka namun itu adalah nilai kategorik, jika di proses dengan data seperti itu maka akan membuat model salah dalam melakukan prediksi.\n",
    "\n",
    "**One-Hot:** season, year, weekday, weathersit, holiday  \n",
    "**Binary:** month, day, hr"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {},
   "outputs": [],
   "source": [
    "transformer = ColumnTransformer([\n",
    "    ('One Hot', OneHotEncoder(drop='first', handle_unknown='ignore'),\n",
    "     ['season', 'year', 'weekday', 'weathersit', 'holiday']),\n",
    "    ('Binary', ce.BinaryEncoder(), ['month', 'day', 'hr'])\n",
    "], remainder='passthrough')\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**Simple Model Benchmarking**\n",
    "\n",
    "Pada tahap ini kita akan melakukan proses pencarian skor dari metrik yang telah kita tentukan sebelumnya, algoritma yang digunakan antara lain `LinearRegression`, `Decision Tree`, `Random Forest`, `XGBoost`, dan `LightGBM`. kita juga menggunakan `Standard Scaler` karena merupakan yang paling tepat untuk kasus regresi dengan tidak terlalu banyak outlier. "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[-1.1524926  -1.2056115  -1.21512698 -1.20249011 -1.15304591] TransformedTargetRegressor(func=<ufunc 'log'>, inverse_func=<ufunc 'exp'>,\n",
      "                           regressor=LinearRegression())\n",
      "[-0.06566896 -0.06608374 -0.06986103 -0.06712477 -0.06411202] TransformedTargetRegressor(func=<ufunc 'log'>, inverse_func=<ufunc 'exp'>,\n",
      "                           regressor=LinearRegression())\n",
      "[-0.95046397 -0.97248758 -1.00624919 -0.9603761  -0.94434092] TransformedTargetRegressor(func=<ufunc 'log'>, inverse_func=<ufunc 'exp'>,\n",
      "                           regressor=LinearRegression())\n",
      "[-0.30188527 -0.28868215 -0.31110902 -0.28943809 -0.29301515] TransformedTargetRegressor(func=<ufunc 'log'>, inverse_func=<ufunc 'exp'>,\n",
      "                           regressor=LinearRegression())\n",
      "[-0.79074958 -0.96293934 -0.80373512 -0.83004952 -0.56101809] TransformedTargetRegressor(func=<ufunc 'log'>, inverse_func=<ufunc 'exp'>,\n",
      "                           regressor=DecisionTreeRegressor(random_state=42))\n",
      "[-0.03262646 -0.04725272 -0.03293269 -0.03604152 -0.01923818] TransformedTargetRegressor(func=<ufunc 'log'>, inverse_func=<ufunc 'exp'>,\n",
      "                           regressor=DecisionTreeRegressor(random_state=42))\n",
      "[-0.50612229 -0.58950356 -0.52068046 -0.5047565  -0.36879451] TransformedTargetRegressor(func=<ufunc 'log'>, inverse_func=<ufunc 'exp'>,\n",
      "                           regressor=DecisionTreeRegressor(random_state=42))\n",
      "[-0.15382392 -0.18316907 -0.16252891 -0.14862209 -0.11497777] TransformedTargetRegressor(func=<ufunc 'log'>, inverse_func=<ufunc 'exp'>,\n",
      "                           regressor=DecisionTreeRegressor(random_state=42))\n",
      "[-0.55600615 -0.65712401 -0.57737359 -0.54054395 -0.40205164] TransformedTargetRegressor(func=<ufunc 'log'>, inverse_func=<ufunc 'exp'>,\n",
      "                           regressor=RandomForestRegressor(random_state=42))\n",
      "[-0.01569089 -0.0217206  -0.01735299 -0.01421394 -0.00933494] TransformedTargetRegressor(func=<ufunc 'log'>, inverse_func=<ufunc 'exp'>,\n",
      "                           regressor=RandomForestRegressor(random_state=42))\n",
      "[-0.41791573 -0.45889029 -0.40861001 -0.38547037 -0.29257693] TransformedTargetRegressor(func=<ufunc 'log'>, inverse_func=<ufunc 'exp'>,\n",
      "                           regressor=RandomForestRegressor(random_state=42))\n",
      "[-0.12325614 -0.1354517  -0.12528604 -0.11202457 -0.08953042] TransformedTargetRegressor(func=<ufunc 'log'>, inverse_func=<ufunc 'exp'>,\n",
      "                           regressor=RandomForestRegressor(random_state=42))\n",
      "[-0.52903214 -0.49625818 -0.4896561  -0.47469556 -0.43026402] TransformedTargetRegressor(func=<ufunc 'log'>, inverse_func=<ufunc 'exp'>,\n",
      "                           regressor=XGBRegressor(base_score=None, booster=None,\n",
      "                                                  colsample_bylevel=None,\n",
      "                                                  colsample_bynode=None,\n",
      "                                                  colsample_bytree=None,\n",
      "                                                  enable_categorical=False,\n",
      "                                                  gamma=None, gpu_id=None,\n",
      "                                                  importance_type=None,\n",
      "                                                  interaction_constraints=None,\n",
      "                                                  learning_rate=None,\n",
      "                                                  max_delta_step=None,\n",
      "                                                  max_depth=None,\n",
      "                                                  min_child_weight=None,\n",
      "                                                  missing=nan,\n",
      "                                                  monotone_constraints=None,\n",
      "                                                  n_estimators=100, n_jobs=None,\n",
      "                                                  num_parallel_tree=None,\n",
      "                                                  predictor=None,\n",
      "                                                  random_state=42,\n",
      "                                                  reg_alpha=None,\n",
      "                                                  reg_lambda=None,\n",
      "                                                  scale_pos_weight=None,\n",
      "                                                  subsample=None,\n",
      "                                                  tree_method=None,\n",
      "                                                  validate_parameters=None,\n",
      "                                                  verbosity=None))\n",
      "[-0.0137111  -0.01280472 -0.01312407 -0.01132774 -0.00989273] TransformedTargetRegressor(func=<ufunc 'log'>, inverse_func=<ufunc 'exp'>,\n",
      "                           regressor=XGBRegressor(base_score=None, booster=None,\n",
      "                                                  colsample_bylevel=None,\n",
      "                                                  colsample_bynode=None,\n",
      "                                                  colsample_bytree=None,\n",
      "                                                  enable_categorical=False,\n",
      "                                                  gamma=None, gpu_id=None,\n",
      "                                                  importance_type=None,\n",
      "                                                  interaction_constraints=None,\n",
      "                                                  learning_rate=None,\n",
      "                                                  max_delta_step=None,\n",
      "                                                  max_depth=None,\n",
      "                                                  min_child_weight=None,\n",
      "                                                  missing=nan,\n",
      "                                                  monotone_constraints=None,\n",
      "                                                  n_estimators=100, n_jobs=None,\n",
      "                                                  num_parallel_tree=None,\n",
      "                                                  predictor=None,\n",
      "                                                  random_state=42,\n",
      "                                                  reg_alpha=None,\n",
      "                                                  reg_lambda=None,\n",
      "                                                  scale_pos_weight=None,\n",
      "                                                  subsample=None,\n",
      "                                                  tree_method=None,\n",
      "                                                  validate_parameters=None,\n",
      "                                                  verbosity=None))\n",
      "[-0.39602867 -0.36727507 -0.36547595 -0.35282423 -0.32011903] TransformedTargetRegressor(func=<ufunc 'log'>, inverse_func=<ufunc 'exp'>,\n",
      "                           regressor=XGBRegressor(base_score=None, booster=None,\n",
      "                                                  colsample_bylevel=None,\n",
      "                                                  colsample_bynode=None,\n",
      "                                                  colsample_bytree=None,\n",
      "                                                  enable_categorical=False,\n",
      "                                                  gamma=None, gpu_id=None,\n",
      "                                                  importance_type=None,\n",
      "                                                  interaction_constraints=None,\n",
      "                                                  learning_rate=None,\n",
      "                                                  max_delta_step=None,\n",
      "                                                  max_depth=None,\n",
      "                                                  min_child_weight=None,\n",
      "                                                  missing=nan,\n",
      "                                                  monotone_constraints=None,\n",
      "                                                  n_estimators=100, n_jobs=None,\n",
      "                                                  num_parallel_tree=None,\n",
      "                                                  predictor=None,\n",
      "                                                  random_state=42,\n",
      "                                                  reg_alpha=None,\n",
      "                                                  reg_lambda=None,\n",
      "                                                  scale_pos_weight=None,\n",
      "                                                  subsample=None,\n",
      "                                                  tree_method=None,\n",
      "                                                  validate_parameters=None,\n",
      "                                                  verbosity=None))\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[-0.11375985 -0.10579548 -0.11041103 -0.10247363 -0.09399373] TransformedTargetRegressor(func=<ufunc 'log'>, inverse_func=<ufunc 'exp'>,\n",
      "                           regressor=XGBRegressor(base_score=None, booster=None,\n",
      "                                                  colsample_bylevel=None,\n",
      "                                                  colsample_bynode=None,\n",
      "                                                  colsample_bytree=None,\n",
      "                                                  enable_categorical=False,\n",
      "                                                  gamma=None, gpu_id=None,\n",
      "                                                  importance_type=None,\n",
      "                                                  interaction_constraints=None,\n",
      "                                                  learning_rate=None,\n",
      "                                                  max_delta_step=None,\n",
      "                                                  max_depth=None,\n",
      "                                                  min_child_weight=None,\n",
      "                                                  missing=nan,\n",
      "                                                  monotone_constraints=None,\n",
      "                                                  n_estimators=100, n_jobs=None,\n",
      "                                                  num_parallel_tree=None,\n",
      "                                                  predictor=None,\n",
      "                                                  random_state=42,\n",
      "                                                  reg_alpha=None,\n",
      "                                                  reg_lambda=None,\n",
      "                                                  scale_pos_weight=None,\n",
      "                                                  subsample=None,\n",
      "                                                  tree_method=None,\n",
      "                                                  validate_parameters=None,\n",
      "                                                  verbosity=None))\n",
      "[-0.4827058  -0.51164513 -0.4466123  -0.46155093 -0.40363532] TransformedTargetRegressor(func=<ufunc 'log'>, inverse_func=<ufunc 'exp'>,\n",
      "                           regressor=LGBMRegressor(random_state=42))\n",
      "[-0.01201701 -0.0129714  -0.01171159 -0.01041804 -0.00901388] TransformedTargetRegressor(func=<ufunc 'log'>, inverse_func=<ufunc 'exp'>,\n",
      "                           regressor=LGBMRegressor(random_state=42))\n",
      "[-0.36863704 -0.37618143 -0.33612504 -0.34586712 -0.30292686] TransformedTargetRegressor(func=<ufunc 'log'>, inverse_func=<ufunc 'exp'>,\n",
      "                           regressor=LGBMRegressor(random_state=42))\n",
      "[-0.10813333 -0.1065739  -0.10481602 -0.09940038 -0.09081612] TransformedTargetRegressor(func=<ufunc 'log'>, inverse_func=<ufunc 'exp'>,\n",
      "                           regressor=LGBMRegressor(random_state=42))\n"
     ]
    }
   ],
   "source": [
    "lr = LinearRegression()\n",
    "dt = DecisionTreeRegressor(random_state=42)\n",
    "rf = RandomForestRegressor(random_state=42)\n",
    "xgb = XGBRegressor(random_state=42)\n",
    "lgb = LGBMRegressor(random_state=42)\n",
    "\n",
    "# Pemodelan dilakukan dalam skala logaritmik, namun kemudian di-inverse kembali untuk interpretasi\n",
    "log_lr = TransformedTargetRegressor(lr, func=np.log, inverse_func=np.exp)\n",
    "log_dt = TransformedTargetRegressor(dt, func=np.log, inverse_func=np.exp)\n",
    "log_rf = TransformedTargetRegressor(rf, func=np.log, inverse_func=np.exp)\n",
    "log_xgb = TransformedTargetRegressor(xgb, func=np.log, inverse_func=np.exp)\n",
    "log_lgb = TransformedTargetRegressor(lgb, func=np.log, inverse_func=np.exp)\n",
    "\n",
    "# Scaling data\n",
    "scaler = StandardScaler()\n",
    "\n",
    "# Kandidat algoritma\n",
    "models = [log_lr, log_dt, log_rf, log_xgb, log_lgb]\n",
    "\n",
    "score_rmse = []\n",
    "nilai_mean_rmse = []\n",
    "nilai_std_rmse = []\n",
    "\n",
    "score_msle = []\n",
    "nilai_mean_msle = []\n",
    "nilai_std_msle = []\n",
    "\n",
    "score_mae = []\n",
    "nilai_mean_mae = []\n",
    "nilai_std_mae = []\n",
    "\n",
    "score_mape = []\n",
    "nilai_mean_mape = []\n",
    "nilai_std_mape = []\n",
    "\n",
    "# Mencari algoritma terbaik berdasarkan nilai MAE\n",
    "for i in models:\n",
    "\n",
    "    crossval = KFold(n_splits=5, shuffle=True, random_state=42)\n",
    "\n",
    "    estimator = Pipeline([\n",
    "        ('dates', DateTransformer()),\n",
    "        ('preprocessing', transformer),\n",
    "        ('scaler', scaler),\n",
    "        ('model', i)\n",
    "    ])\n",
    "    \n",
    "    # RMSE\n",
    "    model_cv_rmse = cross_val_score(\n",
    "        estimator, \n",
    "        X_train, \n",
    "        y_train, \n",
    "        cv=crossval, \n",
    "        scoring='neg_root_mean_squared_error', \n",
    "        error_score='raise'\n",
    "        )\n",
    "\n",
    "    print(model_cv_rmse, i)\n",
    "\n",
    "    score_rmse.append(model_cv_rmse)\n",
    "    nilai_mean_rmse.append(model_cv_rmse.mean())\n",
    "    nilai_std_rmse.append(model_cv_rmse.std())\n",
    "    \n",
    "    # RMSLE - Root Mean Squared Log Error.\n",
    "    # This metric is used when the Target variable is converted into Log(Target).\n",
    "    # This metric reduces errors generated by outliers in dataset\n",
    "    model_cv_msle = cross_val_score(\n",
    "        estimator,\n",
    "        X_train,\n",
    "        y_train,\n",
    "        cv=crossval,\n",
    "        scoring='neg_mean_squared_log_error',\n",
    "        error_score='raise'\n",
    "    )\n",
    "\n",
    "    print(model_cv_msle, i)\n",
    "\n",
    "    score_msle.append(model_cv_msle)\n",
    "    nilai_mean_msle.append(model_cv_msle.mean())\n",
    "    nilai_std_msle.append(model_cv_msle.std())\n",
    "\n",
    "    # MAE\n",
    "    model_cv_mae = cross_val_score(\n",
    "        estimator,\n",
    "        X_train,\n",
    "        y_train,\n",
    "        cv=crossval,\n",
    "        scoring='neg_mean_absolute_error',\n",
    "        error_score='raise'\n",
    "    )\n",
    "\n",
    "    print(model_cv_mae, i)\n",
    "\n",
    "    score_mae.append(model_cv_mae)\n",
    "    nilai_mean_mae.append(model_cv_mae.mean())\n",
    "    nilai_std_mae.append(model_cv_mae.std())\n",
    "\n",
    "    # MAPE\n",
    "    model_cv_mape = cross_val_score(\n",
    "        estimator,\n",
    "        X_train,\n",
    "        y_train,\n",
    "        cv=crossval,\n",
    "        scoring='neg_mean_absolute_percentage_error',\n",
    "        error_score='raise'\n",
    "    )\n",
    "\n",
    "    print(model_cv_mape, i)\n",
    "\n",
    "    score_mape.append(model_cv_mape)\n",
    "    nilai_mean_mape.append(model_cv_mape.mean())\n",
    "    nilai_std_mape.append(model_cv_mape.std())\n"
   ]
  },
  {
   "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>Model</th>\n",
       "      <th>Mean_RMSE</th>\n",
       "      <th>Std_RMSE</th>\n",
       "      <th>Mean_MSLE</th>\n",
       "      <th>Std_MSLE</th>\n",
       "      <th>Mean_MAE</th>\n",
       "      <th>Std_MAE</th>\n",
       "      <th>Mean_MAPE</th>\n",
       "      <th>Std_MAPE</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>Linear Regressor</td>\n",
       "      <td>-1.185753</td>\n",
       "      <td>0.027252</td>\n",
       "      <td>-0.066570</td>\n",
       "      <td>0.001909</td>\n",
       "      <td>-0.966784</td>\n",
       "      <td>0.021914</td>\n",
       "      <td>-0.296826</td>\n",
       "      <td>0.008544</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>DecisionTree Regressor</td>\n",
       "      <td>-0.789698</td>\n",
       "      <td>0.129718</td>\n",
       "      <td>-0.033618</td>\n",
       "      <td>0.008944</td>\n",
       "      <td>-0.497971</td>\n",
       "      <td>0.071684</td>\n",
       "      <td>-0.152624</td>\n",
       "      <td>0.022207</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>RandomForest Regressor</td>\n",
       "      <td>-0.546620</td>\n",
       "      <td>0.082682</td>\n",
       "      <td>-0.015663</td>\n",
       "      <td>0.004042</td>\n",
       "      <td>-0.392693</td>\n",
       "      <td>0.055402</td>\n",
       "      <td>-0.117110</td>\n",
       "      <td>0.015669</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>XGBoost Regressor</td>\n",
       "      <td>-0.483981</td>\n",
       "      <td>0.032200</td>\n",
       "      <td>-0.012172</td>\n",
       "      <td>0.001384</td>\n",
       "      <td>-0.360345</td>\n",
       "      <td>0.024586</td>\n",
       "      <td>-0.105287</td>\n",
       "      <td>0.006837</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>LightGBM Regressor</td>\n",
       "      <td>-0.461230</td>\n",
       "      <td>0.036149</td>\n",
       "      <td>-0.011226</td>\n",
       "      <td>0.001375</td>\n",
       "      <td>-0.345947</td>\n",
       "      <td>0.025985</td>\n",
       "      <td>-0.101948</td>\n",
       "      <td>0.006297</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                    Model  Mean_RMSE  Std_RMSE  Mean_MSLE  Std_MSLE  Mean_MAE  \\\n",
       "0        Linear Regressor  -1.185753  0.027252  -0.066570  0.001909 -0.966784   \n",
       "1  DecisionTree Regressor  -0.789698  0.129718  -0.033618  0.008944 -0.497971   \n",
       "2  RandomForest Regressor  -0.546620  0.082682  -0.015663  0.004042 -0.392693   \n",
       "3       XGBoost Regressor  -0.483981  0.032200  -0.012172  0.001384 -0.360345   \n",
       "4      LightGBM Regressor  -0.461230  0.036149  -0.011226  0.001375 -0.345947   \n",
       "\n",
       "    Std_MAE  Mean_MAPE  Std_MAPE  \n",
       "0  0.021914  -0.296826  0.008544  \n",
       "1  0.071684  -0.152624  0.022207  \n",
       "2  0.055402  -0.117110  0.015669  \n",
       "3  0.024586  -0.105287  0.006837  \n",
       "4  0.025985  -0.101948  0.006297  "
      ]
     },
     "execution_count": 21,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Evaluasi hasil dari 5 kandidat algoritma yang digunakan\n",
    "pd.DataFrame({\n",
    "    'Model': ['Linear Regressor', 'DecisionTree Regressor', 'RandomForest Regressor', 'XGBoost Regressor', 'LightGBM Regressor'],\n",
    "    'Mean_RMSE': nilai_mean_rmse,\n",
    "    'Std_RMSE': nilai_std_rmse,\n",
    "    'Mean_MSLE': nilai_mean_msle,\n",
    "    'Std_MSLE': nilai_std_msle,\n",
    "    'Mean_MAE': nilai_mean_mae,\n",
    "    'Std_MAE': nilai_std_mae,\n",
    "    'Mean_MAPE': nilai_mean_mape,\n",
    "    'Std_MAPE': nilai_std_mape\n",
    "})"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "berdasarkan data diatas, menunjukan terdapat 3 model terbaik berdasarkan neg-MSLE dan metrik lain diatas yaitu `Random Forest`, `XGBoost`, `LightGBM`. sedangkan 2 sisanya terpaut jauh dari ketiga model yang telah disebutkan. selanjutnya, ketiga model terbaik ini akan kita uji lagi dengan melakukan prediksi pada test set.\n",
    "\n",
    "*perlu diperhatikan disini kita masih menggunakan MSLE dan bukan RMSLE dikarenakan tidak tersedianya metrik tersebut pada fungsi sklearn saat penelitian ini dibuat, sedangkan jika ingin di lakukan secara manual menggunakan fungsi square root akan menghasilkan nilai nan karena tidak dapat melakukan perhitungan nilai negatif."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "dict_keys(['explained_variance', 'r2', 'max_error', 'neg_median_absolute_error', 'neg_mean_absolute_error', 'neg_mean_absolute_percentage_error', 'neg_mean_squared_error', 'neg_mean_squared_log_error', 'neg_root_mean_squared_error', 'neg_mean_poisson_deviance', 'neg_mean_gamma_deviance', 'accuracy', 'top_k_accuracy', 'roc_auc', 'roc_auc_ovr', 'roc_auc_ovo', 'roc_auc_ovr_weighted', 'roc_auc_ovo_weighted', 'balanced_accuracy', 'average_precision', 'neg_log_loss', 'neg_brier_score', 'adjusted_rand_score', 'rand_score', 'homogeneity_score', 'completeness_score', 'v_measure_score', 'mutual_info_score', 'adjusted_mutual_info_score', 'normalized_mutual_info_score', 'fowlkes_mallows_score', 'precision', 'precision_macro', 'precision_micro', 'precision_samples', 'precision_weighted', 'recall', 'recall_macro', 'recall_micro', 'recall_samples', 'recall_weighted', 'f1', 'f1_macro', 'f1_micro', 'f1_samples', 'f1_weighted', 'jaccard', 'jaccard_macro', 'jaccard_micro', 'jaccard_samples', 'jaccard_weighted'])"
      ]
     },
     "execution_count": 22,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "from sklearn import metrics\n",
    "\n",
    "metrics.SCORERS.keys()\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**Simple Modeling and Predicting with RandomForest, XGBoost, LightGBM**"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "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>RMSE</th>\n",
       "      <th>RMSLE</th>\n",
       "      <th>MAE</th>\n",
       "      <th>MAPE</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>Random Forest</th>\n",
       "      <td>159.989316</td>\n",
       "      <td>1.160036</td>\n",
       "      <td>109.834887</td>\n",
       "      <td>2.043436</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>XGB</th>\n",
       "      <td>79.931778</td>\n",
       "      <td>0.460243</td>\n",
       "      <td>47.352365</td>\n",
       "      <td>0.427581</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>LightGBM</th>\n",
       "      <td>86.048454</td>\n",
       "      <td>0.524619</td>\n",
       "      <td>52.473721</td>\n",
       "      <td>0.519813</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                     RMSE     RMSLE         MAE      MAPE\n",
       "Random Forest  159.989316  1.160036  109.834887  2.043436\n",
       "XGB             79.931778  0.460243   47.352365  0.427581\n",
       "LightGBM        86.048454  0.524619   52.473721  0.519813"
      ]
     },
     "execution_count": 23,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "models = {\n",
    "    'Random Forest': lr,\n",
    "    'XGB': xgb,\n",
    "    'LightGBM': lgb\n",
    "}\n",
    "\n",
    "score_rmse = []\n",
    "score_rmsle = []\n",
    "score_mae = []\n",
    "score_mape = []\n",
    "\n",
    "# Prediksi pada test set\n",
    "for i in models:\n",
    "\n",
    "    model = Pipeline([\n",
    "        ('dates', DateTransformer()),\n",
    "        ('preprocessing', transformer),\n",
    "        ('scaler', scaler),\n",
    "        ('model', models[i])\n",
    "    ])\n",
    "\n",
    "    model.fit(X_train, y_train)\n",
    "    y_pred = model.predict(X_test)\n",
    "    \n",
    "    score_rmse.append(np.sqrt(mean_squared_error(np.expm1(y_test), np.expm1(y_pred))))\n",
    "    score_rmsle.append(np.sqrt(mean_squared_log_error(np.expm1(y_test), np.expm1(y_pred))))\n",
    "    score_mae.append(mean_absolute_error(np.expm1(y_test), np.expm1(y_pred)))\n",
    "    score_mape.append(mean_absolute_percentage_error(np.expm1(y_test), np.expm1(y_pred)))\n",
    "\n",
    "score_before_tuning = pd.DataFrame(\n",
    "    {'RMSE': score_rmse, 'RMSLE': score_rmsle, 'MAE': score_mae, 'MAPE': score_mape}, index=models.keys())\n",
    "score_before_tuning"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "ketika melakukan prediksi pada test set disini terlihat bahwa dari segi RMSLE, maupun skor lain, `XGBoost` merupakan yang terbaik, model 'LightGBM' juga termasuk baik disini, tetapi karena target kita transformasi menggunakan `Logarithm Transformation` maka kita akan menggunakan skor RMSLE sebagai acuan utamanya. namun sayangnya pada saat penelitian ini dibuat berdasarkan [Dokumentasi LightGBM](https://lightgbm.readthedocs.io/en/latest/Parameters.html#metric-parameters), algoritma `LightGBM` ini tidak mempunyai metrics tersebut. oleh karena itu kita hanya akan menggunakan algoritma terbaik  berdasarkan data diatas yaitu `XGBoost` untuk kita tunning pada langkah selanjutnya.\n",
    "\n",
    "*Perlu diperhatikan, sebelumnya kita telah melakukan transformasi mengunakan logaritma pada kolom target yaitu `cnt` oleh karena itu, melakukan skoring menggunakan data tersebut merupakan hal yang kurang tepat, maka disini kita menggunakan fungsi np.expm1() yang merupakan inverse dari np.log1p() yang kita gunakan sebelumnya yang berarti dapat digunakan untuk mengembalikan nilai aslinya, dan hasilnya kita jadi dapat membandingkan skor pada nilai prediksi dan aktual dengan skala yang sebenarnya."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**Hyperparameter Tuning**\n",
    "\n",
    "Karena pada model XGBoost kita mendapatkan performa terbaik dan dapat menggunaakan metrik RMSLE, maka selanjutnya kita kan menggunakan model ini hingga akhir penelitian. pada tahap ini, kita akan mencoba meningkatkan performa model dengan melakukan hyperparameter tuning. \n",
    "\n",
    "disini kita akan melakukan hyperparameter tuning menggunakan `RandomizedSearchCV` dikarenakan kinerjanya yang sangat ringan walaupun memiliki kekurangan yaitu proses tuning dilakukan secara acak dan tidak menyeluruh seperti yang dilakukan `GridSearchCV`\n",
    "\n",
    "untuk panduan tuning untuk model XGBoost, dapat mengacu pada dokuemntasinya di [XGBoost Parameters](https://xgboost.readthedocs.io/en/stable/parameter.html)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Kedalaman pohon\n",
    "max_depth = list(np.arange(1, 16))\n",
    "\n",
    "# Learning rate\n",
    "# list(np.arange(0.01, 1.01, 0.01))\n",
    "learning_rate = list(np.arange(1, 100)/100)\n",
    "\n",
    "# Jumlah pohon\n",
    "n_estimators = list(np.arange(100, 201))\n",
    "\n",
    "# Jumlah baris tiap pohon (% dari total baris train set)\n",
    "subsample = list(np.arange(2, 10)/10)\n",
    "\n",
    "# Gamma (min_impurity_decrease)\n",
    "# Semakin besar nilainya, semakin konservatif/simpel modelnya\n",
    "gamma = list(np.arange(1, 11))\n",
    "\n",
    "# Jumlah feature yang digunakan untuk tiap pohon (% dari total kolom train set)\n",
    "colsample_bytree = list(np.arange(1, 10)/10)\n",
    "\n",
    "# Alpha (regularization)\n",
    "# Semakin besar nilainya, semakin konservatif/simpel modelnya\n",
    "reg_alpha = list(np.logspace(-3, 1, 10))  # [0.01, 0.1, 1, 10, 100]\n",
    "\n",
    "\n",
    "# Hyperparam space XGboost\n",
    "hyperparam_space = {\n",
    "    'model__max_depth': max_depth,\n",
    "    'model__learning_rate': learning_rate,\n",
    "    'model__n_estimators': n_estimators,\n",
    "    'model__subsample': subsample,\n",
    "    'model__gamma': gamma,\n",
    "    'model__colsample_bytree': colsample_bytree,\n",
    "    'model__reg_alpha': reg_alpha\n",
    "}\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Benchmark model dengan hyperparameter tuning\n",
    "xgb = XGBRegressor(random_state=42, verbosity=0)\n",
    "\n",
    "# Membuat algorithm chains\n",
    "estimator_xgb = Pipeline([\n",
    "    ('dates', DateTransformer()),\n",
    "    ('preprocessing', transformer),\n",
    "    ('scaler', scaler),\n",
    "    ('model', xgb)\n",
    "])\n",
    "\n",
    "crossval = KFold(n_splits=5, shuffle=True, random_state=42)\n",
    "\n",
    "# Hyperparameter tuning\n",
    "random_search = RandomizedSearchCV(\n",
    "    estimator_xgb,\n",
    "    param_distributions=hyperparam_space,\n",
    "    n_iter=50,\n",
    "    cv=crossval,\n",
    "    scoring=['neg_root_mean_squared_error', 'neg_mean_squared_log_error', \n",
    "             'neg_mean_absolute_error','neg_mean_absolute_percentage_error'],\n",
    "    n_jobs=-1,\n",
    "    # Hanya bisa memilih salah stau metric untuk optimisasi\n",
    "    refit='neg_mean_squared_log_error',\n",
    "    random_state=42,\n",
    "    verbose=1\n",
    ")\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Fitting 5 folds for each of 50 candidates, totalling 250 fits\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "RandomizedSearchCV(cv=KFold(n_splits=5, random_state=42, shuffle=True),\n",
       "                   estimator=Pipeline(steps=[('dates', DateTransformer()),\n",
       "                                             ('preprocessing',\n",
       "                                              ColumnTransformer(remainder='passthrough',\n",
       "                                                                transformers=[('One '\n",
       "                                                                               'Hot',\n",
       "                                                                               OneHotEncoder(drop='first',\n",
       "                                                                                             handle_unknown='ignore'),\n",
       "                                                                               ['season',\n",
       "                                                                                'year',\n",
       "                                                                                'weekday',\n",
       "                                                                                'weathersit',\n",
       "                                                                                'holiday']),\n",
       "                                                                              ('Binary',\n",
       "                                                                               BinaryEncoder(),\n",
       "                                                                               ['month',\n",
       "                                                                                'da...\n",
       "                                                             0.021544346900318832,\n",
       "                                                             0.05994842503189409,\n",
       "                                                             0.1668100537200059,\n",
       "                                                             0.46415888336127775,\n",
       "                                                             1.2915496650148828,\n",
       "                                                             3.593813663804626,\n",
       "                                                             10.0],\n",
       "                                        'model__subsample': [0.2, 0.3, 0.4, 0.5,\n",
       "                                                             0.6, 0.7, 0.8,\n",
       "                                                             0.9]},\n",
       "                   random_state=42, refit='neg_mean_squared_log_error',\n",
       "                   scoring=['neg_root_mean_squared_error',\n",
       "                            'neg_mean_squared_log_error',\n",
       "                            'neg_mean_absolute_error',\n",
       "                            'neg_mean_absolute_percentage_error'],\n",
       "                   verbose=1)"
      ]
     },
     "execution_count": 26,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "random_search.fit(X_train, y_train)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "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>mean_fit_time</th>\n",
       "      <th>std_fit_time</th>\n",
       "      <th>mean_score_time</th>\n",
       "      <th>std_score_time</th>\n",
       "      <th>param_model__subsample</th>\n",
       "      <th>param_model__reg_alpha</th>\n",
       "      <th>param_model__n_estimators</th>\n",
       "      <th>param_model__max_depth</th>\n",
       "      <th>param_model__learning_rate</th>\n",
       "      <th>param_model__gamma</th>\n",
       "      <th>...</th>\n",
       "      <th>std_test_neg_mean_absolute_error</th>\n",
       "      <th>rank_test_neg_mean_absolute_error</th>\n",
       "      <th>split0_test_neg_mean_absolute_percentage_error</th>\n",
       "      <th>split1_test_neg_mean_absolute_percentage_error</th>\n",
       "      <th>split2_test_neg_mean_absolute_percentage_error</th>\n",
       "      <th>split3_test_neg_mean_absolute_percentage_error</th>\n",
       "      <th>split4_test_neg_mean_absolute_percentage_error</th>\n",
       "      <th>mean_test_neg_mean_absolute_percentage_error</th>\n",
       "      <th>std_test_neg_mean_absolute_percentage_error</th>\n",
       "      <th>rank_test_neg_mean_absolute_percentage_error</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>28</th>\n",
       "      <td>10.554282</td>\n",
       "      <td>0.084779</td>\n",
       "      <td>0.040622</td>\n",
       "      <td>0.007649</td>\n",
       "      <td>0.8</td>\n",
       "      <td>0.002783</td>\n",
       "      <td>186</td>\n",
       "      <td>11</td>\n",
       "      <td>0.32</td>\n",
       "      <td>4</td>\n",
       "      <td>...</td>\n",
       "      <td>0.014470</td>\n",
       "      <td>1</td>\n",
       "      <td>-0.102189</td>\n",
       "      <td>-0.103345</td>\n",
       "      <td>-0.105384</td>\n",
       "      <td>-0.105460</td>\n",
       "      <td>-0.093540</td>\n",
       "      <td>-0.101983</td>\n",
       "      <td>0.004401</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>10</th>\n",
       "      <td>3.684343</td>\n",
       "      <td>0.137426</td>\n",
       "      <td>0.049996</td>\n",
       "      <td>0.015311</td>\n",
       "      <td>0.9</td>\n",
       "      <td>0.059948</td>\n",
       "      <td>109</td>\n",
       "      <td>8</td>\n",
       "      <td>0.48</td>\n",
       "      <td>5</td>\n",
       "      <td>...</td>\n",
       "      <td>0.020955</td>\n",
       "      <td>2</td>\n",
       "      <td>-0.107854</td>\n",
       "      <td>-0.112635</td>\n",
       "      <td>-0.109338</td>\n",
       "      <td>-0.109269</td>\n",
       "      <td>-0.097667</td>\n",
       "      <td>-0.107352</td>\n",
       "      <td>0.005091</td>\n",
       "      <td>2</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>14</th>\n",
       "      <td>18.787310</td>\n",
       "      <td>0.331218</td>\n",
       "      <td>0.040620</td>\n",
       "      <td>0.007652</td>\n",
       "      <td>0.9</td>\n",
       "      <td>0.007743</td>\n",
       "      <td>164</td>\n",
       "      <td>15</td>\n",
       "      <td>0.46</td>\n",
       "      <td>10</td>\n",
       "      <td>...</td>\n",
       "      <td>0.014225</td>\n",
       "      <td>3</td>\n",
       "      <td>-0.113344</td>\n",
       "      <td>-0.114989</td>\n",
       "      <td>-0.114983</td>\n",
       "      <td>-0.110072</td>\n",
       "      <td>-0.106875</td>\n",
       "      <td>-0.112053</td>\n",
       "      <td>0.003150</td>\n",
       "      <td>3</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>3.516737</td>\n",
       "      <td>0.097025</td>\n",
       "      <td>0.040620</td>\n",
       "      <td>0.007655</td>\n",
       "      <td>0.8</td>\n",
       "      <td>0.002783</td>\n",
       "      <td>175</td>\n",
       "      <td>6</td>\n",
       "      <td>0.66</td>\n",
       "      <td>6</td>\n",
       "      <td>...</td>\n",
       "      <td>0.022028</td>\n",
       "      <td>4</td>\n",
       "      <td>-0.131784</td>\n",
       "      <td>-0.128729</td>\n",
       "      <td>-0.128029</td>\n",
       "      <td>-0.113759</td>\n",
       "      <td>-0.121700</td>\n",
       "      <td>-0.124800</td>\n",
       "      <td>0.006421</td>\n",
       "      <td>4</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9</th>\n",
       "      <td>10.885545</td>\n",
       "      <td>0.472753</td>\n",
       "      <td>0.040626</td>\n",
       "      <td>0.007654</td>\n",
       "      <td>0.9</td>\n",
       "      <td>0.464159</td>\n",
       "      <td>162</td>\n",
       "      <td>12</td>\n",
       "      <td>0.14</td>\n",
       "      <td>2</td>\n",
       "      <td>...</td>\n",
       "      <td>0.034202</td>\n",
       "      <td>5</td>\n",
       "      <td>-0.136075</td>\n",
       "      <td>-0.141503</td>\n",
       "      <td>-0.127559</td>\n",
       "      <td>-0.163477</td>\n",
       "      <td>-0.143340</td>\n",
       "      <td>-0.142391</td>\n",
       "      <td>0.011885</td>\n",
       "      <td>5</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>5 rows × 44 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "    mean_fit_time  std_fit_time  mean_score_time  std_score_time  \\\n",
       "28      10.554282      0.084779         0.040622        0.007649   \n",
       "10       3.684343      0.137426         0.049996        0.015311   \n",
       "14      18.787310      0.331218         0.040620        0.007652   \n",
       "0        3.516737      0.097025         0.040620        0.007655   \n",
       "9       10.885545      0.472753         0.040626        0.007654   \n",
       "\n",
       "   param_model__subsample param_model__reg_alpha param_model__n_estimators  \\\n",
       "28                    0.8               0.002783                       186   \n",
       "10                    0.9               0.059948                       109   \n",
       "14                    0.9               0.007743                       164   \n",
       "0                     0.8               0.002783                       175   \n",
       "9                     0.9               0.464159                       162   \n",
       "\n",
       "   param_model__max_depth param_model__learning_rate param_model__gamma  ...  \\\n",
       "28                     11                       0.32                  4  ...   \n",
       "10                      8                       0.48                  5  ...   \n",
       "14                     15                       0.46                 10  ...   \n",
       "0                       6                       0.66                  6  ...   \n",
       "9                      12                       0.14                  2  ...   \n",
       "\n",
       "   std_test_neg_mean_absolute_error rank_test_neg_mean_absolute_error  \\\n",
       "28                         0.014470                                 1   \n",
       "10                         0.020955                                 2   \n",
       "14                         0.014225                                 3   \n",
       "0                          0.022028                                 4   \n",
       "9                          0.034202                                 5   \n",
       "\n",
       "    split0_test_neg_mean_absolute_percentage_error  \\\n",
       "28                                       -0.102189   \n",
       "10                                       -0.107854   \n",
       "14                                       -0.113344   \n",
       "0                                        -0.131784   \n",
       "9                                        -0.136075   \n",
       "\n",
       "    split1_test_neg_mean_absolute_percentage_error  \\\n",
       "28                                       -0.103345   \n",
       "10                                       -0.112635   \n",
       "14                                       -0.114989   \n",
       "0                                        -0.128729   \n",
       "9                                        -0.141503   \n",
       "\n",
       "    split2_test_neg_mean_absolute_percentage_error  \\\n",
       "28                                       -0.105384   \n",
       "10                                       -0.109338   \n",
       "14                                       -0.114983   \n",
       "0                                        -0.128029   \n",
       "9                                        -0.127559   \n",
       "\n",
       "    split3_test_neg_mean_absolute_percentage_error  \\\n",
       "28                                       -0.105460   \n",
       "10                                       -0.109269   \n",
       "14                                       -0.110072   \n",
       "0                                        -0.113759   \n",
       "9                                        -0.163477   \n",
       "\n",
       "    split4_test_neg_mean_absolute_percentage_error  \\\n",
       "28                                       -0.093540   \n",
       "10                                       -0.097667   \n",
       "14                                       -0.106875   \n",
       "0                                        -0.121700   \n",
       "9                                        -0.143340   \n",
       "\n",
       "    mean_test_neg_mean_absolute_percentage_error  \\\n",
       "28                                     -0.101983   \n",
       "10                                     -0.107352   \n",
       "14                                     -0.112053   \n",
       "0                                      -0.124800   \n",
       "9                                      -0.142391   \n",
       "\n",
       "    std_test_neg_mean_absolute_percentage_error  \\\n",
       "28                                     0.004401   \n",
       "10                                     0.005091   \n",
       "14                                     0.003150   \n",
       "0                                      0.006421   \n",
       "9                                      0.011885   \n",
       "\n",
       "    rank_test_neg_mean_absolute_percentage_error  \n",
       "28                                             1  \n",
       "10                                             2  \n",
       "14                                             3  \n",
       "0                                              4  \n",
       "9                                              5  \n",
       "\n",
       "[5 rows x 44 columns]"
      ]
     },
     "execution_count": 27,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Melihat hasil tuning dalam bentuk dataframe. Diurutkan berdasarkan RMSLE, MAE & MAPE\n",
    "pd.DataFrame(random_search.cv_results_).sort_values(by=['rank_test_neg_mean_squared_log_error',\n",
    "                                                     'rank_test_neg_mean_absolute_error', 'rank_test_neg_mean_absolute_percentage_error']).head()\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "XGBoost\n",
      "Best_score: -0.01157409004906427\n",
      "Best_params: {'model__subsample': 0.8, 'model__reg_alpha': 0.0027825594022071257, 'model__n_estimators': 186, 'model__max_depth': 11, 'model__learning_rate': 0.32, 'model__gamma': 4, 'model__colsample_bytree': 0.9}\n"
     ]
    }
   ],
   "source": [
    "# Cek skor dan params terbaik\n",
    "print('XGBoost')\n",
    "print('Best_score:', random_search.best_score_)\n",
    "print('Best_params:', random_search.best_params_)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**Tunned Modeling and Predicting with XGBoost**\n",
    "\n",
    "Tahap ini melakukan proses fitting ulang menggunaakn parameter terbaik yang didapat dari tahap tuning sebelumnya, lalu melakukan prediksi pada test set."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Model XGBoost\n",
    "model = {'XGB': XGBRegressor(random_state=42)}\n",
    "\n",
    "# Define model terhadap estimator terbaik\n",
    "xgb_tuning = random_search.best_estimator_\n",
    "\n",
    "# Fitting model\n",
    "xgb_tuning.fit(X_train, y_train)\n",
    "\n",
    "# Predict test set\n",
    "y_pred_xgb_tuning = xgb_tuning.predict(X_test)\n",
    "\n",
    "# Simpan nilai metrics RMSE, RMSLE, MAE & MAPE setelah tuning\n",
    "rmse_xgb_tuning = np.sqrt(mean_squared_error(np.expm1(y_test), np.expm1(y_pred_xgb_tuning)))\n",
    "rmsle_xgb_tuning = np.sqrt(mean_squared_log_error(np.expm1(y_test), np.expm1(y_pred_xgb_tuning)))\n",
    "mae_xgb_tuning = mean_absolute_error(np.expm1(y_test), np.expm1(y_pred_xgb_tuning))\n",
    "mape_xgb_tuning = mean_absolute_percentage_error(np.expm1(y_test), np.expm1(y_pred_xgb_tuning))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**Performance Comparison**"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "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>RMSE</th>\n",
       "      <th>RMSLE</th>\n",
       "      <th>MAE</th>\n",
       "      <th>MAPE</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>XGB Before Tuning</th>\n",
       "      <td>79.931778</td>\n",
       "      <td>0.460243</td>\n",
       "      <td>47.352365</td>\n",
       "      <td>0.427581</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>XGB After Tuning</th>\n",
       "      <td>77.962755</td>\n",
       "      <td>0.471320</td>\n",
       "      <td>47.121819</td>\n",
       "      <td>0.437451</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                        RMSE     RMSLE        MAE      MAPE\n",
       "XGB Before Tuning  79.931778  0.460243  47.352365  0.427581\n",
       "XGB After Tuning   77.962755  0.471320  47.121819  0.437451"
      ]
     },
     "execution_count": 30,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "final_score = {'XGB Before Tuning': {'RMSE': score_before_tuning.loc['XGB']['RMSE'],\n",
    "                              'RMSLE': score_before_tuning.loc['XGB']['RMSLE'],\n",
    "                              'MAE': score_before_tuning.loc['XGB']['MAE'],\n",
    "                              'MAPE': score_before_tuning.loc['XGB']['MAPE']},\n",
    "               'XGB After Tuning': {'RMSE': rmse_xgb_tuning, 'RMSLE': rmsle_xgb_tuning, \n",
    "                                    'MAE': mae_xgb_tuning, 'MAPE': mape_xgb_tuning},}\n",
    "pd.DataFrame(final_score).T"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Terlihat bahwa skor RMSE dan MAE mengalami sedikit kenaikan namun untuk RMSLE dan MAPE malah mengalami sangat sedikit penurunan setelah melakukan hyperparameter tuning"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 31,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 864x504 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 864x504 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.figure(figsize=(12, 7))\n",
    "plot = sns.scatterplot(x=np.expm1(y_test), y=np.expm1(y_pred)).set(title='Actual vs. Prediction Price Before Tuning',\n",
    "                                                                   xlabel='Actual Bike Use',\n",
    "                                                                   ylabel='Predicted Bike Use')\n",
    "plt.show()\n",
    "\n",
    "plt.figure(figsize=(12, 7))\n",
    "plot = sns.scatterplot(x=np.expm1(y_test), y=np.expm1(y_pred_xgb_tuning)).set(title='Actual vs. Prediction Price After Tuning',\n",
    "                                                          xlabel='Actual Bike Use',\n",
    "                                                          ylabel='Predicted Bike Use')\n",
    "plt.show()\n",
    "\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "berdasarkan kedua grafik diatas, terlihat bahwa semakin besar nilainya maka semakin besar juga kemungkinan terjadi error, dimana ini merupakan suatu limitasi dari dataset yang kita gunakan yang hanya berjumlah sekitar 12 ribu saja. data kebanyakan tersebar di sejumlah kecil sepeda yang digunakan, hanya sedikit kondisi dimana sepeda sangat banyak digunakan, hal ini membuat model gagal memprediksi dengan baik kondisi tersebut dimana sepeda banyak digunakan."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**Feature Importances**\n",
    "\n",
    "Untuk dapat mengetahui fitur apa saja yang paling memengaruhi target, kita dapat mengeceknya menggunakan function feature_importances_."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 32,
   "metadata": {},
   "outputs": [],
   "source": [
    "estimator_xgb = Pipeline([('dates', DateTransformer()),\n",
    "                        ('preprocessing', transformer),\n",
    "                        ('scaler', scaler),\n",
    "                        ('model', xgb)\n",
    "                    ])\n",
    "\n",
    "estimator_xgb.fit(X_train, y_train);"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 33,
   "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"
    },
    {
     "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": [
    "\n",
    "feature_imp_before = pd.Series(estimator_xgb['model'].feature_importances_, transformer.get_feature_names()).sort_values(ascending = False).head()\n",
    "feature_imp_after = pd.Series(xgb_tuning['model'].feature_importances_, transformer.get_feature_names()).sort_values(ascending = False).head()\n",
    "\n",
    "\n",
    "feature_imp_before.plot(kind='barh', title='Feature Importances Before Tuning');\n",
    "plt.show()\n",
    "feature_imp_after.plot(kind='barh', title='Feature Importances After Tuning');\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**Predicted Vs Actual Value**\n",
    "\n",
    "pada tahap ini kita ingin membandingkan nilai aktual dengan hasil prediksi menggunakan model sebelum dituning dan setelah dituning "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 34,
   "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>prediction</th>\n",
       "      <th>tunned prediction</th>\n",
       "      <th>actual</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>1958</th>\n",
       "      <td>459.079002</td>\n",
       "      <td>537.781189</td>\n",
       "      <td>481.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>10536</th>\n",
       "      <td>297.803294</td>\n",
       "      <td>383.620331</td>\n",
       "      <td>404.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7010</th>\n",
       "      <td>202.372078</td>\n",
       "      <td>180.067459</td>\n",
       "      <td>180.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4133</th>\n",
       "      <td>2.263715</td>\n",
       "      <td>2.085660</td>\n",
       "      <td>3.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>364</th>\n",
       "      <td>24.349680</td>\n",
       "      <td>17.618425</td>\n",
       "      <td>11.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6385</th>\n",
       "      <td>133.145289</td>\n",
       "      <td>116.269012</td>\n",
       "      <td>92.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5492</th>\n",
       "      <td>17.275450</td>\n",
       "      <td>23.406687</td>\n",
       "      <td>18.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3910</th>\n",
       "      <td>163.898831</td>\n",
       "      <td>115.481438</td>\n",
       "      <td>243.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>8730</th>\n",
       "      <td>74.891492</td>\n",
       "      <td>71.289940</td>\n",
       "      <td>37.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7817</th>\n",
       "      <td>135.228278</td>\n",
       "      <td>129.291443</td>\n",
       "      <td>93.0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>2400 rows × 3 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "       prediction  tunned prediction  actual\n",
       "1958   459.079002         537.781189   481.0\n",
       "10536  297.803294         383.620331   404.0\n",
       "7010   202.372078         180.067459   180.0\n",
       "4133     2.263715           2.085660     3.0\n",
       "364     24.349680          17.618425    11.0\n",
       "...           ...                ...     ...\n",
       "6385   133.145289         116.269012    92.0\n",
       "5492    17.275450          23.406687    18.0\n",
       "3910   163.898831         115.481438   243.0\n",
       "8730    74.891492          71.289940    37.0\n",
       "7817   135.228278         129.291443    93.0\n",
       "\n",
       "[2400 rows x 3 columns]"
      ]
     },
     "execution_count": 34,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Because of model programmed with log, then we neet to change back the prediction to real value with np.expm1()\n",
    "pred_tuned = np.expm1(y_pred_xgb_tuning)\n",
    "pred = np.expm1(y_pred)\n",
    "actual = np.expm1(y_test) \n",
    "\n",
    "validation = pd.DataFrame({'prediction': pred, 'tunned prediction': pred_tuned, 'actual': actual})\n",
    "validation\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### **Conclusion**\n",
    "---"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Berdasarkan pemodelan yang telah kita lakukan dapat disimpulkan bahwa model yang kita gunakan memiliki kecenderungan semakin tinggi nilainya maka akan semakin menyebar, hal ini memungkinkan karena terdapat limitasi pada dataset yang memiliki kecenderungan serupa.\n",
    "\n",
    "Dari segi fitur yang paling berpengaruh untuk model sebelum tuning masing-masing adalah `hr`, dan `workingday`, sedangkan pada model setelah tuning memiliki fitur yang paling berpengaruh yang berbeda yaitu `hr`, `temp`, dan `atemp`.\n",
    "\n",
    "Berdasarkan metrik yang digunakan pada penelitian ini yaitu RMSE, RMSLE, MAE, dan MAPE. jika dilihat dari hasilnya, nilai RMSE setelah tuning memiliki skor yang lebih baik hal ini membuat prediksi yang di dapat semakin akurat berdasarkan rataannya, tetapi jika disandingkan dengan RMSLE nilai yang dihasilkan sebelum melakukan hyperparameter tuning mempunyai RMSE sebesar 79.9, dengan RMSLE sebesar 0.460, sedangkan untuk nilai setelah melakukan hyperparameter tunning mempunyai RMSE sebesar 77.96, dengan RMSLE sebesar 0.47. eror yang dihasilkan RMSLE pada model sebelum tuning justru lebih kecil dengan RMSE nya yang justru lebih besar, hal ini terjadi karena RMSLE merupakan suatu rasio dimana ketika persebaran yang ada lebih besar maka metrik malah akan membuat erornya lebih kecil, hal ini berguna jika kita ingin menyamaratakan prediksi yang menghasilkan nilai terlalu tinggi atau terlalu rendah. metrik RMSLE cukup tepat digunakan pada kasus persediaan yang biasanya memang terjadi perbedaan pada tiap kondisi dan lokasi.\n",
    "\n",
    "untuk MAE dan MAPE sendiri mirip seperti RMSE dan RMSLE, untuk perbedaannya MAE mengambil nilai absolutnya, sedangkan RMSE mengambil nilai kudaratiknya. untuk MAPE sendiri dihasilkan dari presentase nilai eror yang dihasilkan dari MAE."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### **Recommendation**\n",
    "---"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Untuk mengembangkan model agar dapat menjadi lebih baik lagi dapat dilakukan hal-hal berikut ini:\n",
    "1. Mengecek performa pada model lain yang tidak kita gunakan saat ini seperti Ridge, Lasso, ataupun yang lebih kompleks seperti Neural Network, dll. hal ini dapat meningkatkan kemungkinan adanya model yang berpotensi menjadi lebih baik lagi. terutama pada model LightGBM yang tidak jadi dipakai karena masalah kompatibilitas, bisa dibuatkan fungsi untuk pengecekan performa model agar dapat mengeluarkan metrik dari RMSLE yang telah kita jadikan acuan utama pada penelitian ini.\n",
    "\n",
    "1. Lakukan penambahan fitur yang mungkin akan mempunyai korelasi tinggi terhadap target, misalkan seperti lokasi, kecepatan angin, perkiraan cuaca beberapa waktu kedepan, lama sepeda digunakan, dll. jika memungkinkan, lakukan penambahan data secara signifikan dengan mungkin memasukan data dari wilayan lain yang tidak tercakup pada data yang kita gunakan, hal ini perlu dilakukan untuk menambah keakuratan model yang dihasilkan.\n",
    "\n",
    "1. Optimalkan parameter yang digunakan pada hyperparameter tuning, seperti yang telah diteliti sebelumnya bahwa hyperparameter tuning yang kita lakukan malah membuat performa model menurun, hal ini terjadi karena adanya limitasi pada pengetahuan yang dimiliki sehingga parameter yang digunakan hanya berdasarkan 'trial and error', diharapkan kedepannya dapat meningkatkan pengetahuannya pada hal tersebut.\n",
    "\n",
    "1. Dataset yang digunakan memungkinkan kita untuk melakukan 'anomaly detection', yaitu melakukan pendeteksian jika terjadi hal diluar kebiasaan, misal seperti jam 1 malam yang seharusnya hanya sedikit atau mungkin tidak ada sama sekali yang menyewa sepeda namun yang terjadi saat itu adalah sebaliknya, sepeda yang disewa pada saat itu malah sangat tinggi yang mungkin disebabkan oleh faktor eksternal seperti misalnya adanya konser ataupun acara besar lainnya didekat lokasi penyewaan sepeda pada waktu tersebut, hal ini membuat sistem mendeteksi adanya kebiasaan yang tidak wajar, fungsinya sendiri karena penyewaan sepeda ini dilakukan secara online dan terintegrasi maka penyedia penyewaan sepeda dapat secara cepat melakukan distribusi pada area tersebut jika memungkinkan."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### **Save, Load, Predict Model with Pickle**\n",
    "---"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**Save Model**"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Untuk model akhir yang digunakan adalah model hasil setelah tuning karena memiliki RMSE yang lebih kecil walaupun dengan RMSLE yang lebih besar, hal ini dikarenakan asumsi sebelumnya yang menyatakan RSMLE baik digunakan untuk menyamaratakan hasil prediksi yang nilainya jauh berbeda dan tepat digunakan pada model yang mempunyai banyak perbedaan karena suatu alasan misal perbedaan kebiasaan di negara berbeda dan lainnya, sedangkan data yang kita gunakan hanya bertempat di satu lokasi dengan hanya melihat kondisi dan waktunya saja."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 35,
   "metadata": {},
   "outputs": [],
   "source": [
    "filename = 'Model_BikeSharing_XGB.sav'\n",
    "\n",
    "# Jika menggunakan model sebelum tuning\n",
    "# estimator = Pipeline([('dates', DateTransformer()),\n",
    "#                       ('preprocessing', transformer),\n",
    "#                       ('scaler', scaler),\n",
    "#                       ('model', xgb)])\n",
    "# estimator.fit(X_train, y_train)\n",
    "# pickle.dump(estimator, open(filename, 'wb'))\n",
    "\n",
    "# Jika Menggunakan Model Setelah di Tuning\n",
    "pickle.dump(xgb_tuning, open(filename, 'wb'))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**Load Model**"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Don't forget to run libraries section\n",
    "loaded_model = pickle.load(open(filename, 'rb'))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**Predict New Data**"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Read New Dataset\n",
    "# test = pd.read_csv()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "# USE THIS ONLY IF WANT TO MEASURE ERROR WITH LOGARITHM TRANSFORMATION ON CNT FEATURE\n",
    "# NOT RECOMMENDED, BECAUSE IT WILL MEASURE WITHOUT THE ACTUAL VALUE\n",
    "# test['cnt'] = test['cnt'].apply(lambda x: np.log1p(x)) \n",
    "\n",
    "X_new = test.drop(['cnt'], axis=1)\n",
    "y_new = test['cnt']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "y_pred_test = loaded_model.predict(X_new)\n",
    "\n",
    "def new_model(scoring):\n",
    "    return scoring(y_new, np.expm1(y_pred_test)) # IF NOT USE LOGARITHN TRANSFORMATION\n",
    "    # return scoring(y_new, y_pred_test) # IF USE LOGARITHM TRANSFORMATION\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "print('RMSLE:', np.sqrt(new_model(mean_squared_log_error)))\n",
    "print('MAE:', new_model(mean_absolute_error))\n",
    "print('MAPE:', new_model(mean_absolute_percentage_error))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Because of model programmed with log, then we neet to change back the prediction to real value with np.expm1()\n",
    "pred_test = np.expm1(y_pred_test) \n",
    "actual_test = y_new  # IF NOT USE LOGARITHM TRANSFORMATION\n",
    "# actual_test = np.expm1(y_new)  # IF USE LOGARITHM TRANSFORMATION\n",
    "\n",
    "prediction = pd.DataFrame({'prediction': pred_test, 'actual': actual_test})\n",
    "\n",
    "prediction['prediction'] = prediction.apply(lambda x: 1 if (x['prediction'] <= 0) else x['prediction'], axis=1)\n",
    "# prediction.to_csv(\"prediction.csv\", index=False)\n",
    "\n",
    "prediction\n"
   ]
  }
 ],
 "metadata": {
  "interpreter": {
   "hash": "5294d7ba68cfa59cd3a94a139d76e7f1c79534bbcd875ba49e2f30f535f55ba1"
  },
  "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": "3.9.12"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}