house_prices/day25-submit-to-kaggle.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "DAY 25 -- Mar 21, 2017"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "import pandas as pd\n",
    "import matplotlib.pyplot as plt\n",
    "import seaborn as sns\n",
    "import os\n",
    "\n",
    "%matplotlib inline"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "from sklearn.linear_model import LinearRegression\n",
    "from sklearn.model_selection import cross_val_score"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "input_dir = \"./data/\"\n",
    "\n",
    "# Load my data\n",
    "train_file = os.path.join(input_dir, \"train.csv\")\n",
    "test_file = os.path.join(input_dir, \"test.csv\")\n",
    "\n",
    "train = pd.read_csv(train_file)\n",
    "test = pd.read_csv(test_file)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "y = train[[\"SalePrice\"]]\n",
    "X = train.drop([\"SalePrice\", \"Id\"], axis=1)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Features"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(1460, 79)\n"
     ]
    }
   ],
   "source": [
    "# Number of features\n",
    "print(X.shape)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {
    "collapsed": false,
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Number of categorical features: 43 \n",
      "\n",
      "MSZoning:\t ['RL' 'RM' 'C (all)' 'FV' 'RH']\n",
      "Street:\t ['Pave' 'Grvl']\n",
      "Alley:\t [nan 'Grvl' 'Pave']\n",
      "LotShape:\t ['Reg' 'IR1' 'IR2' 'IR3']\n",
      "LandContour:\t ['Lvl' 'Bnk' 'Low' 'HLS']\n",
      "Utilities:\t ['AllPub' 'NoSeWa']\n",
      "LotConfig:\t ['Inside' 'FR2' 'Corner' 'CulDSac' 'FR3']\n",
      "LandSlope:\t ['Gtl' 'Mod' 'Sev']\n",
      "Neighborhood:\t ['CollgCr' 'Veenker' 'Crawfor' 'NoRidge' 'Mitchel' 'Somerst' 'NWAmes'\n",
      " 'OldTown' 'BrkSide' 'Sawyer' 'NridgHt' 'NAmes' 'SawyerW' 'IDOTRR'\n",
      " 'MeadowV' 'Edwards' 'Timber' 'Gilbert' 'StoneBr' 'ClearCr' 'NPkVill'\n",
      " 'Blmngtn' 'BrDale' 'SWISU' 'Blueste']\n",
      "Condition1:\t ['Norm' 'Feedr' 'PosN' 'Artery' 'RRAe' 'RRNn' 'RRAn' 'PosA' 'RRNe']\n",
      "Condition2:\t ['Norm' 'Artery' 'RRNn' 'Feedr' 'PosN' 'PosA' 'RRAn' 'RRAe']\n",
      "BldgType:\t ['1Fam' '2fmCon' 'Duplex' 'TwnhsE' 'Twnhs']\n",
      "HouseStyle:\t ['2Story' '1Story' '1.5Fin' '1.5Unf' 'SFoyer' 'SLvl' '2.5Unf' '2.5Fin']\n",
      "RoofStyle:\t ['Gable' 'Hip' 'Gambrel' 'Mansard' 'Flat' 'Shed']\n",
      "RoofMatl:\t ['CompShg' 'WdShngl' 'Metal' 'WdShake' 'Membran' 'Tar&Grv' 'Roll' 'ClyTile']\n",
      "Exterior1st:\t ['VinylSd' 'MetalSd' 'Wd Sdng' 'HdBoard' 'BrkFace' 'WdShing' 'CemntBd'\n",
      " 'Plywood' 'AsbShng' 'Stucco' 'BrkComm' 'AsphShn' 'Stone' 'ImStucc'\n",
      " 'CBlock']\n",
      "Exterior2nd:\t ['VinylSd' 'MetalSd' 'Wd Shng' 'HdBoard' 'Plywood' 'Wd Sdng' 'CmentBd'\n",
      " 'BrkFace' 'Stucco' 'AsbShng' 'Brk Cmn' 'ImStucc' 'AsphShn' 'Stone' 'Other'\n",
      " 'CBlock']\n",
      "MasVnrType:\t ['BrkFace' 'None' 'Stone' 'BrkCmn' nan]\n",
      "ExterQual:\t ['Gd' 'TA' 'Ex' 'Fa']\n",
      "ExterCond:\t ['TA' 'Gd' 'Fa' 'Po' 'Ex']\n",
      "Foundation:\t ['PConc' 'CBlock' 'BrkTil' 'Wood' 'Slab' 'Stone']\n",
      "BsmtQual:\t ['Gd' 'TA' 'Ex' nan 'Fa']\n",
      "BsmtCond:\t ['TA' 'Gd' nan 'Fa' 'Po']\n",
      "BsmtExposure:\t ['No' 'Gd' 'Mn' 'Av' nan]\n",
      "BsmtFinType1:\t ['GLQ' 'ALQ' 'Unf' 'Rec' 'BLQ' nan 'LwQ']\n",
      "BsmtFinType2:\t ['Unf' 'BLQ' nan 'ALQ' 'Rec' 'LwQ' 'GLQ']\n",
      "Heating:\t ['GasA' 'GasW' 'Grav' 'Wall' 'OthW' 'Floor']\n",
      "HeatingQC:\t ['Ex' 'Gd' 'TA' 'Fa' 'Po']\n",
      "CentralAir:\t ['Y' 'N']\n",
      "Electrical:\t ['SBrkr' 'FuseF' 'FuseA' 'FuseP' 'Mix' nan]\n",
      "KitchenQual:\t ['Gd' 'TA' 'Ex' 'Fa']\n",
      "Functional:\t ['Typ' 'Min1' 'Maj1' 'Min2' 'Mod' 'Maj2' 'Sev']\n",
      "FireplaceQu:\t [nan 'TA' 'Gd' 'Fa' 'Ex' 'Po']\n",
      "GarageType:\t ['Attchd' 'Detchd' 'BuiltIn' 'CarPort' nan 'Basment' '2Types']\n",
      "GarageFinish:\t ['RFn' 'Unf' 'Fin' nan]\n",
      "GarageQual:\t ['TA' 'Fa' 'Gd' nan 'Ex' 'Po']\n",
      "GarageCond:\t ['TA' 'Fa' nan 'Gd' 'Po' 'Ex']\n",
      "PavedDrive:\t ['Y' 'N' 'P']\n",
      "PoolQC:\t [nan 'Ex' 'Fa' 'Gd']\n",
      "Fence:\t [nan 'MnPrv' 'GdWo' 'GdPrv' 'MnWw']\n",
      "MiscFeature:\t [nan 'Shed' 'Gar2' 'Othr' 'TenC']\n",
      "SaleType:\t ['WD' 'New' 'COD' 'ConLD' 'ConLI' 'CWD' 'ConLw' 'Con' 'Oth']\n",
      "SaleCondition:\t ['Normal' 'Abnorml' 'Partial' 'AdjLand' 'Alloca' 'Family']\n"
     ]
    }
   ],
   "source": [
    "categorical_var = X.dtypes[X.dtypes == object]\n",
    "\n",
    "print(\"Number of categorical features:\", len(categorical_var), \"\\n\")\n",
    "for f in categorical_var.index:\n",
    "    print(\"{}:\\t {}\".format(f, pd.unique(X[f])))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {
    "collapsed": false,
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(1460, 288)\n",
      "(1460, 285)\n"
     ]
    }
   ],
   "source": [
    "# Number of features after mapping of categorical features to numbers\n",
    "X2 = pd.get_dummies(X)\n",
    "print(X2.shape)\n",
    "\n",
    "# Drop na\n",
    "X2.dropna(axis=1, how='any', inplace=True)\n",
    "print(X2.shape)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Kaggle: Linear regression of only quantitative variables"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "33"
      ]
     },
     "execution_count": 8,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Filter for only quantitative columns\n",
    "X_quant = X[(X.dtypes[X.dtypes!=object]).index]\n",
    "X_quant = X_quant.dropna(axis=1)\n",
    "X_quant.shape\n",
    "\n",
    "features = X_quant.columns\n",
    "len(features)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "# Ensure test has features\n",
    "features_test = test[features].dropna(axis=1).columns\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "25"
      ]
     },
     "execution_count": 10,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Unify features\n",
    "features = list(set(features).intersection(set(features_test)))\n",
    "len(features)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "The scores: [ 0.83387649  0.77871025  0.69953849]\n",
      "Accuracy: 0.77 (+/- 0.11)\n"
     ]
    }
   ],
   "source": [
    "clf = LinearRegression()\n",
    "scores = cross_val_score(clf, X_quant[features], y, cv=3)\n",
    "\n",
    "# Print scores\n",
    "print(\"The scores:\", scores)\n",
    "print(\"Accuracy: %0.2f (+/- %0.2f)\" % (scores.mean(), scores.std() * 2))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[1]\n"
     ]
    }
   ],
   "source": [
    "clf = LinearRegression()\n",
    "clf.fit(X_quant[features], y)\n",
    "y_test = clf.predict(test[features])\n",
    "\n",
    "# Check that is not negative\n",
    "print(sum(y_test < 0))\n",
    "y_test[y_test<0] = 0\n",
    "\n",
    "# Submission\n",
    "predictions = pd.concat([test[\"Id\"], pd.DataFrame(y_test, columns=[\"SalePrice\"])], axis=1)\n",
    "predictions.to_csv(\"saleprice_submission-linreg-quant25.csv\", index=False)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "KAGGLE: 0.43803"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Kaggle linear regression of one-hot + quantitative features"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "The scores: [ 0.87766846  0.82330935  0.73942924]\n",
      "Accuracy: 0.81 (+/- 0.11)\n"
     ]
    }
   ],
   "source": [
    "clf = LinearRegression()\n",
    "scores = cross_val_score(clf, X2, y, cv=3)\n",
    "\n",
    "# Print scores\n",
    "print(\"The scores:\", scores)\n",
    "print(\"Accuracy: %0.2f (+/- %0.2f)\" % (scores.mean(), scores.std() * 2))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "259\n"
     ]
    }
   ],
   "source": [
    "# Transform test set\n",
    "test_dummies = pd.get_dummies(test).dropna(axis=1, how='any')\n",
    "common_features = list(set(X2.columns).intersection(set(test_dummies.columns)))\n",
    "common_features = [f for f in common_features if f != \"Id\"]\n",
    "print(len(common_features))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "The scores: [ 0.85407967  0.80668306  0.74405846]\n",
      "Accuracy: 0.80 (+/- 0.09)\n"
     ]
    }
   ],
   "source": [
    "clf = LinearRegression()\n",
    "scores = cross_val_score(clf, X2[common_features], y, cv=3)\n",
    "\n",
    "# Print scores\n",
    "print(\"The scores:\", scores)\n",
    "print(\"Accuracy: %0.2f (+/- %0.2f)\" % (scores.mean(), scores.std() * 2))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[1]\n"
     ]
    }
   ],
   "source": [
    "# Make prediction\n",
    "clf = LinearRegression()\n",
    "clf.fit(X2[common_features], y)\n",
    "y_test = clf.predict(test_dummies[common_features])\n",
    "\n",
    "# Check that is not negative\n",
    "print(sum(y_test < 0))\n",
    "y_test[y_test<0] = 0\n",
    "\n",
    "# Submission\n",
    "predictions = pd.concat([test[\"Id\"], pd.DataFrame(y_test, columns=[\"SalePrice\"])], axis=1)\n",
    "predictions.to_csv(\"saleprice_submission-linreg-all259.csv\", index=False)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "KAGGLE: Your submission scored 0.45646"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.axes._subplots.AxesSubplot at 0x1179b19b0>"
      ]
     },
     "execution_count": 17,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAYUAAAD3CAYAAADyvkg2AAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x117965908>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "predictions[[\"SalePrice\"]].plot()"
   ]
  }
 ],
 "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": "3.6.0"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}