03_Modeling.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Modeling"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Libraries to Import"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "import pandas as pd\n",
    "import numpy as np\n",
    "import Builder\n",
    "\n",
    "from sklearn.dummy import DummyRegressor\n",
    "from sklearn.model_selection import train_test_split\n",
    "from sklearn.tree import DecisionTreeRegressor\n",
    "\n",
    "from sklearn.metrics import mean_squared_error as mse\n",
    "from sklearn.metrics import r2_score"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Read in Data\n",
    "\n",
    "This is a fairly common dataset, here sourced from [data.world](https://data.world/data-society/capital-bikeshare-2011-2012), but also available from the UCI database. It details hundreds of thousands of Capital Bikeshare rides in 2011 and 2012 sampled every hour, including features for temporal factors, weather conditions, and registered versus casual users. "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "df = pd.read_csv('Data/bike_data.csv')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Data Cleaning"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Our date column was an object, so transform to a datetime with pandas and create a new column for Month."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "df.Date = pd.to_datetime(arg=df.Date) #dtype('<M8[ns]')\n",
    "df['Month'] = df['Date'].dt.month"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Remove Correlated Features\n",
    "\n",
    "We'll be predicting Total Users, so let's drop the other user columns because they'll be heavily correlated with Total Users. And we'll remove Temperatuer F, in favor of the \"feels like\" temperature. "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [],
   "source": [
    "df.drop(['Casual Users', 'Registered Users', 'Temperature F', 'Date'], axis=1, inplace=True)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Modeling"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Encode Categorical Data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [],
   "source": [
    "cate_feats = ['Season', 'Hour', 'Holiday', 'Day of the Week', 'Working Day', 'Weather Type', 'Month']\n",
    "\n",
    "for col in cate_feats:\n",
    "    df[col] = df[col].astype('category')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [],
   "source": [
    "X = pd.get_dummies(df, drop_first=True)\n",
    "X.drop('Total Users', axis=1, inplace=True)\n",
    "y = df['Total Users']"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Normalization"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [],
   "source": [
    "for col in X.columns:\n",
    "    # Subtract the minimum and divide by the range forcing a scale of 0 to 1 for each feature\n",
    "    X[col] = (X[col] - min(X[col])) / (max(X[col]) - min(X[col])) "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Train-Test-Split "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [],
   "source": [
    "X_train, X_test, y_train, y_test = train_test_split(X, y, random_state=10)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Dummy Regressor"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Dummy Model Accuracy:  0.0\n"
     ]
    }
   ],
   "source": [
    "dummy_regressor = DummyRegressor(strategy='mean')\n",
    "dummy_regressor.fit(X, y)\n",
    "dummy_regressor.predict(X)\n",
    "print(\"Dummy Model Accuracy: \", dummy_regressor.score(X, y))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### CART Tree Model"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "DecisionTreeRegressor(max_depth=3, random_state=42)"
      ]
     },
     "execution_count": 10,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "regressor = DecisionTreeRegressor(random_state=42, max_depth=3)\n",
    "regressor.fit(X_train, y_train)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "MSE score: 21091.847432693474\n",
      "R-sq score: 0.344608475006546\n"
     ]
    }
   ],
   "source": [
    "y_pred = regressor.predict(X_test)\n",
    "print('MSE score:', mse(y_test, y_pred))\n",
    "print('R-sq score:', r2_score(y_test,y_pred))"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.7.6"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 4
}