solution.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Case study of the Risk Assesment GmbH"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "import os\n",
    "import pandas as pd"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Encoding problems"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "The following lines in the file **RAcredit_train.csv** contain encripted characters yielding encoding errors:\n",
    "- 43240\n",
    "- 44795\n",
    "- 48316\n",
    "\n",
    "To avoid these errors, we drop the column named **emp_title**."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "def load_data(name, nrows):\n",
    "    badnames = ['emp_title']\n",
    "    goodcols = lambda x: x not in badnames\n",
    "    df = pd.read_csv(name, usecols=goodcols, nrows=nrows, low_memory=False)\n",
    "    return df"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "path = \"/home/jose/Learning/cases\"\n",
    "df = load_data(os.path.join(path, \"RAcredit_train.csv\"), None)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(77450, 135)"
      ]
     },
     "execution_count": 4,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df.shape"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Before cleaning, the train dataset contains **77450** rows and **135** columns."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Handle missing data"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Drop rows with null *l_state*"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(76809, 135)"
      ]
     },
     "execution_count": 5,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df = df[df.l_state.notnull()]\n",
    "df.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Fully Paid    74163\n",
       "Default        2646\n",
       "Name: l_state, dtype: int64"
      ]
     },
     "execution_count": 6,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df.l_state.value_counts()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Drop columns with any missing value"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(76809, 53)"
      ]
     },
     "execution_count": 7,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df = df.dropna(axis=1, how='any')\n",
    "df.shape"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "After cleaning, the train dataset contains **76809** rows and **53** columns."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Analysis of categorical data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "def cat_cols(df):\n",
    "    return set(df.columns) - set(df._get_numeric_data().columns)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "{'addr_state',\n",
       " 'earliest_cr_line',\n",
       " 'grade',\n",
       " 'home_ownership',\n",
       " 'initial_list_status',\n",
       " 'issue_d',\n",
       " 'l_state',\n",
       " 'pymnt_plan',\n",
       " 'sub_grade',\n",
       " 'term',\n",
       " 'verification_status'}"
      ]
     },
     "execution_count": 9,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "cat_cols(df)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Remove categorical data with more than 10 categories"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "for col in cat_cols(df):\n",
    "    n = len(df[col].unique())\n",
    "    if n>10:\n",
    "        df.drop([col], axis=1, inplace=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "{'grade', 'l_state', 'term'}"
      ]
     },
     "execution_count": 11,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "cat_cols(df)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Convert categorical variable into dummy variables"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "df = pd.get_dummies(df, columns=list(cat_cols(df)))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Correlation analysis"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Keep correlated variables with abs(corr)>0.1 "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "corr_matrix = df.corr()\n",
    "corr_matrix = corr_matrix[abs(corr_matrix[\"l_state_Default\"]) > 0.1]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAdUAAAEICAYAAAAJGW4GAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\ngg==\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7fb285905080>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "import matplotlib.pyplot as plt\n",
    "%matplotlib inline\n",
    "\n",
    "corr_matrix.l_state_Default.plot(kind='barh', grid=True, title=\"Correlation Analysis\");"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(76809, 10)"
      ]
     },
     "execution_count": 15,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "good_variables = corr_matrix.l_state_Default.index\n",
    "df = df[good_variables]\n",
    "df.drop(['l_state_Fully Paid'], axis=1, inplace=True)\n",
    "df.shape"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "After the correlation analysis, only 10 columns remain in the dataframe."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Classification models"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "from sklearn.preprocessing import StandardScaler\n",
    "\n",
    "y = df.l_state_Default\n",
    "df.drop(['l_state_Default'], axis=1, inplace=True)\n",
    "X = StandardScaler().fit_transform(df)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "from sklearn.ensemble import RandomForestClassifier\n",
    "model1 = RandomForestClassifier(max_depth=10, random_state=42);\n",
    "\n",
    "from sklearn.neural_network import MLPClassifier\n",
    "model2 = MLPClassifier(solver='lbfgs', hidden_layer_sizes=(10, 10), random_state=42)\n",
    "\n",
    "from sklearn.dummy import DummyClassifier\n",
    "dummy = DummyClassifier(strategy='most_frequent')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAA4gAAAFNCAYAAABCCkHgAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nkgTAVVYAVHwkoDL+AAAAAElFTkSuQmCC\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7fb253483ef0>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "import scikitplot as skplt\n",
    "\n",
    "fig, ax = plt.subplots(1, 3, sharey=True, figsize=(15,5))\n",
    "skplt.estimators.plot_learning_curve(model1, X, y, ax=ax[0], title='Random Forest')\n",
    "skplt.estimators.plot_learning_curve(model2, X, y, ax=ax[1], title='Neural Network')\n",
    "skplt.estimators.plot_learning_curve(dummy, X, y, ax=ax[2], title='Dummy Classifier');"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "The model 1 displays higher score."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "RandomForestClassifier(bootstrap=True, class_weight=None, criterion='gini',\n",
       "            max_depth=10, max_features='auto', max_leaf_nodes=None,\n",
       "            min_impurity_decrease=0.0, min_impurity_split=None,\n",
       "            min_samples_leaf=1, min_samples_split=2,\n",
       "            min_weight_fraction_leaf=0.0, n_estimators=10, n_jobs=1,\n",
       "            oob_score=False, random_state=42, verbose=0, warm_start=False)"
      ]
     },
     "execution_count": 19,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "model1.fit(X, y)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Model predictions"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(19213, 135)"
      ]
     },
     "execution_count": 20,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df = load_data(os.path.join(path, \"RAcredit_test.csv\"), None)\n",
    "df.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(19213, 9)"
      ]
     },
     "execution_count": 21,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df = df[good_variables.drop(['l_state_Default', 'l_state_Fully Paid'])]\n",
    "df.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "X = StandardScaler().fit_transform(df)\n",
    "y_pred = model1.predict(X)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Counter({0: 18996, 1: 217})"
      ]
     },
     "execution_count": 23,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "from collections import Counter\n",
    "\n",
    "Counter(y_pred)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "The model predicts **217 defaults** in the test set."
   ]
  }
 ],
 "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": "3.6.3"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}