demo.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "import pandas as pd\n",
    "import pandas_profiling as pp\n",
    "\n",
    "import seaborn as sns\n",
    "sns.set()\n",
    "import matplotlib.pyplot as plt\n",
    "%matplotlib inline\n",
    "\n",
    "import numpy as np\n",
    "\n",
    "import itertools\n",
    "\n",
    "from sklearn.model_selection import train_test_split\n",
    "from sklearn.metrics import mean_squared_error, r2_score\n",
    "from sklearn.linear_model import Ridge\n",
    "\n",
    "import lightgbm as lgb"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Data exploration"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "df = pd.read_csv(\"data/ames.csv\")\n",
    "df = df[[\"LotArea\", \"OverallQual\", \"OverallCond\", \"YearBuilt\", \"TotalBsmtSF\",\n",
    "         \"1stFlrSF\", \"2ndFlrSF\", \"Fireplaces\", \"GarageCars\", \"OpenPorchSF\", \"SalePrice\"]]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "df.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "df.info()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "df.describe()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "df[\"SalePrice\"].hist()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "sns.clustermap(df.corr());"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "sns.catplot(data=df,\n",
    "            x=\"GarageCars\", y=\"SalePrice\",\n",
    "            col=\"Fireplaces\",\n",
    "            kind=\"bar\",\n",
    "            col_wrap=2);"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "pp.ProfileReport(df).to_file(output_file=\"report.html\")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Regression task"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "X = df.drop(\"SalePrice\", axis=1)\n",
    "Y = df[\"SalePrice\"]"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Train/test split"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "X_train, X_test, Y_train, Y_test = train_test_split(X, Y, test_size=0.25, random_state=12)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "len(X), len(X_train), len(Y_test)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Approach 0: constant model"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "Y_pred_const = np.full(Y_test.shape, Y_train.mean())\n",
    "Y_pred_const[:3]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "np.sqrt(mean_squared_error(Y_test, Y_pred_const))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "r2_score(Y_test, Y_pred_const)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "def plot_predictions(Y_true, Y_predicted):\n",
    "    df_pred = pd.DataFrame({\"true\": Y_true, \"predicted\": Y_predicted})\n",
    "    val_min = 0\n",
    "    val_max = 1000000\n",
    "    figure = sns.jointplot('true', 'predicted', df_pred,\n",
    "                           xlim=(val_min, val_max),\n",
    "                           ylim=(val_min, val_max))\n",
    "    figure.ax_joint.plot([val_min, val_max], [val_min, val_max], ':k') \n",
    "    figure.fig.set_figheight(8)\n",
    "    figure.fig.set_figwidth(8)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "plot_predictions(Y_test, Y_pred_const)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Approach 1: linear regression model"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Preprocessing"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "X_train_lr = X_train.copy()\n",
    "X_test_lr  = X_test.copy()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "X_train_lr[\"LotArea\"] = np.log10(X_train_lr[\"LotArea\"])\n",
    "X_test_lr[\"LotArea\"]  = np.log10(X_test_lr[\"LotArea\"])\n",
    "\n",
    "X_train_lr[\"OpenPorchSF\"] = [np.log10(value) if value > 0 else 0 for value in X_train_lr[\"OpenPorchSF\"]]\n",
    "X_test_lr[\"OpenPorchSF\"]  = [np.log10(value) if value > 0 else 0 for value in X_test_lr[\"OpenPorchSF\"]]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "columns = X_train_lr.columns\n",
    "\n",
    "for col1, col2 in itertools.combinations(columns, 2):\n",
    "    X_train_lr[col1 + '_x_' + col2] = X_train_lr[col1] * X_train_lr[col2]\n",
    "    X_test_lr[col1 + '_x_' + col2]  = X_test_lr[col1] * X_test_lr[col2]\n",
    "    \n",
    "for col in columns:\n",
    "    X_train_lr[col + '^2'] = X_train_lr[col] ** 2\n",
    "    X_test_lr[col + '^2'] = X_test_lr[col] ** 2"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "X_test_lr = (X_test_lr - X_train_lr.mean()) / X_train_lr.std()\n",
    "X_train_lr = (X_train_lr - X_train_lr.mean()) / X_train_lr.std()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "X_train_lr.head()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Training"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "reg = Ridge(random_state=12)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "reg.fit(X_train_lr, Y_train)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "Y_pred_lr = reg.predict(X_test_lr)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "np.sqrt(mean_squared_error(Y_test, Y_pred_lr))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "r2_score(Y_test, Y_pred_lr)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "plot_predictions(Y_test, Y_pred_lr)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Approach 2: LightGBM"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### No preprocessing, just training"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "lgb_model = lgb.LGBMRegressor(random_state=12)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "lgb_model.fit(X_train, Y_train)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "Y_pred_lgb = lgb_model.predict(X_test)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "np.sqrt(mean_squared_error(Y_test, Y_pred_lgb))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "r2_score(Y_test, Y_pred_lgb)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "plot_predictions(Y_test, Y_pred_lgb)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.6.8"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 4
}