Skip to content

Instantly share code, notes, and snippets.

@odubno

odubno/Oleh_Dubno Lending Club Loan Data

Created November 19, 2014 05:22
Show Gist options
  • Save odubno/37bc500ad7554a6d4a59 to your computer and use it in GitHub Desktop.
Save odubno/37bc500ad7554a6d4a59 to your computer and use it in GitHub Desktop.
Download ZIP
Oleh_Dubno Lending Club Loan Data_Draft
Raw
Oleh_Dubno Lending Club Loan Data
{
"metadata": {
"name": "",
"signature": "sha256:c8b47c1d474c58c90ff683c6b4e9f9348a7c6a9e4ad6d8da3ac5f11acc04a112"
},
"nbformat": 3,
"nbformat_minor": 0,
"worksheets": [
{
"cells": [
{
"cell_type": "markdown",
"metadata": {},
"source": [
"#Loan Data (2007-2011) From Lending Club"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"####We will be going over two things:\n",
"\n",
"\u2022 Logistic regression, to determine which features of the data set contribute towards someone paying off their loan or defaulting on their loan in United States.\n",
"\n",
"\u2022 Using cartodb to map the features of the data set and see which states stand out among the rest in terms of paying back their loans."
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"####Here's the data set:\n",
"https://www.lendingclub.com/info/download-data.action"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"##Cleaning the Data"
]
},
{
"cell_type": "code",
"collapsed": false,
"input": [
"import pandas as pd\n",
"import numpy as np\n",
"from datetime import datetime\n",
"from matplotlib import pyplot as plt\n",
"\n",
"%matplotlib inline"
],
"language": "python",
"metadata": {},
"outputs": [],
"prompt_number": 385
},
{
"cell_type": "code",
"collapsed": false,
"input": [
"url = '/Users/olehdubno/Desktop/python_tests/LoanStats3b2.csv'\n",
"loan = pd.read_csv(url, low_memory = False)"
],
"language": "python",
"metadata": {},
"outputs": [],
"prompt_number": 386
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"Creating a separate set of features we will be cleaning and working with."
]
},
{
"cell_type": "code",
"collapsed": false,
"input": [
"loan_2 = loan[['funded_amnt','emp_length','annual_inc','loan_status','home_ownership','addr_state','tax_liens','grade']]"
],
"language": "python",
"metadata": {},
"outputs": [],
"prompt_number": 387
},
{
"cell_type": "code",
"collapsed": false,
"input": [
"loan_2.head()"
],
"language": "python",
"metadata": {},
"outputs": [
{
"html": [
"<div style=\"max-height:1000px;max-width:1500px;overflow:auto;\">\n",
"<table border=\"1\" class=\"dataframe\">\n",
" <thead>\n",
" <tr style=\"text-align: right;\">\n",
" <th></th>\n",
" <th>funded_amnt</th>\n",
" <th>emp_length</th>\n",
" <th>annual_inc</th>\n",
" <th>loan_status</th>\n",
" <th>home_ownership</th>\n",
" <th>addr_state</th>\n",
" <th>tax_liens</th>\n",
" <th>grade</th>\n",
" </tr>\n",
" </thead>\n",
" <tbody>\n",
" <tr>\n",
" <th>0</th>\n",
" <td> 24000</td>\n",
" <td> 10+ years</td>\n",
" <td> 100000</td>\n",
" <td> Current</td>\n",
" <td> MORTGAGE</td>\n",
" <td> MI</td>\n",
" <td> 0</td>\n",
" <td> B</td>\n",
" </tr>\n",
" <tr>\n",
" <th>1</th>\n",
" <td> 11100</td>\n",
" <td> 10+ years</td>\n",
" <td> 90000</td>\n",
" <td> Current</td>\n",
" <td> MORTGAGE</td>\n",
" <td> NY</td>\n",
" <td> 0</td>\n",
" <td> C</td>\n",
" </tr>\n",
" <tr>\n",
" <th>2</th>\n",
" <td> 12000</td>\n",
" <td> 3 years</td>\n",
" <td> 96500</td>\n",
" <td> Current</td>\n",
" <td> MORTGAGE</td>\n",
" <td> TX</td>\n",
" <td> 0</td>\n",
" <td> A</td>\n",
" </tr>\n",
" <tr>\n",
" <th>3</th>\n",
" <td> 15000</td>\n",
" <td> 10+ years</td>\n",
" <td> 98000</td>\n",
" <td> Fully Paid</td>\n",
" <td> RENT</td>\n",
" <td> NY</td>\n",
" <td> 0</td>\n",
" <td> C</td>\n",
" </tr>\n",
" <tr>\n",
" <th>4</th>\n",
" <td> 27600</td>\n",
" <td> 6 years</td>\n",
" <td> 73000</td>\n",
" <td> Current</td>\n",
" <td> MORTGAGE</td>\n",
" <td> CO</td>\n",
" <td> 0</td>\n",
" <td> D</td>\n",
" </tr>\n",
" </tbody>\n",
"</table>\n",
"</div>"
],
"metadata": {},
"output_type": "pyout",
"prompt_number": 388,
"text": [
" funded_amnt emp_length annual_inc loan_status home_ownership addr_state \\\n",
"0 24000 10+ years 100000 Current MORTGAGE MI \n",
"1 11100 10+ years 90000 Current MORTGAGE NY \n",
"2 12000 3 years 96500 Current MORTGAGE TX \n",
"3 15000 10+ years 98000 Fully Paid RENT NY \n",
"4 27600 6 years 73000 Current MORTGAGE CO \n",
"\n",
" tax_liens grade \n",
"0 0 B \n",
"1 0 C \n",
"2 0 A \n",
"3 0 C \n",
"4 0 D "
]
}
],
"prompt_number": 388
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"Cleaning:\n",
"\n",
"\u2022Convert \"loan_status\" to booleans\n",
"\n",
"\u2022Clean up \"emp_length\"\n",
"\n",
"\u2022Convert \"grade\" to integer values\n",
"\n",
"\u2022Drop N/A values from fields"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"Dropping N/A values (It's only 4 rows and not very significant)"
]
},
{
"cell_type": "code",
"collapsed": false,
"input": [
"loan_2.dropna().info()"
],
"language": "python",
"metadata": {},
"outputs": [
{
"output_type": "stream",
"stream": "stdout",
"text": [
"<class 'pandas.core.frame.DataFrame'>\n",
"Int64Index: 188123 entries, 0 to 188122\n",
"Data columns (total 8 columns):\n",
"funded_amnt 188123 non-null float64\n",
"emp_length 188123 non-null object\n",
"annual_inc 188123 non-null float64\n",
"loan_status 188123 non-null object\n",
"home_ownership 188123 non-null object\n",
"addr_state 188123 non-null object\n",
"tax_liens 188123 non-null float64\n",
"grade 188123 non-null object\n",
"dtypes: float64(3), object(5)"
]
}
],
"prompt_number": 389
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"###Cleaning Loan_Status\n",
"\n",
"Our feature \"loan_status\" has seven unique values. To do our logistic regression we require two.\n",
"\n",
"Below, you'll notice a bar chart that highlights each of the seven values. We'll be removing \"Current\", as our goal is to focus on who paid or didn't pay their loans. \"Fully Paid\" will remain as is and the rest of the columns will be characterized as \"Unpaid\", after all that's pretty much what they are."
]
},
{
"cell_type": "code",
"collapsed": false,
"input": [
"loan_2.loan_status.value_counts().plot(kind='bar',alpha=.30)"
],
"language": "python",
"metadata": {},
"outputs": [
{
"metadata": {},
"output_type": "pyout",
"prompt_number": 390,
"text": [
"<matplotlib.axes._subplots.AxesSubplot at 0x120320690>"
]
},
{
"metadata": {},
"output_type": "display_data",
"png": "iVBORw0KGgoAAAANSUhEUgAAAYUAAAFUCAYAAADLQSDCAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\niA3SKURExAbpFCIiYoP/D55gGwZRKhIwAAAAAElFTkSuQmCC\n",
"text": [
"<matplotlib.figure.Figure at 0x1184e73d0>"
]
}
],
"prompt_number": 390
},
{
"cell_type": "code",
"collapsed": false,
"input": [
"#cleaning \"loan_status\"\n",
"loan_2['loan_status_clean'] = loan_2['loan_status'].map({'Current': 2, 'Fully Paid': 1, 'Charged Off':0, 'Late(31-120 days)':0, 'In Grace Period': 0, 'Late(16-30 days)': 0, 'Default': 0})\n",
"loan_2 = loan_2[loan_2.loan_status_clean != 2] \n",
"loan_2[\"loan_status_clean\"] = loan_2[\"loan_status_clean\"].apply(lambda loan_status_clean: 0 if loan_status_clean == 0 else 1)"
],
"language": "python",
"metadata": {},
"outputs": [],
"prompt_number": 391
},
{
"cell_type": "code",
"collapsed": false,
"input": [
"loan_2.loan_status_clean.value_counts().plot(kind='bar',alpha=.30)"
],
"language": "python",
"metadata": {},
"outputs": [
{
"metadata": {},
"output_type": "pyout",
"prompt_number": 392,
"text": [
"<matplotlib.axes._subplots.AxesSubplot at 0x10be29750>"
]
},
{
"metadata": {},
"output_type": "display_data",
"png": "iVBORw0KGgoAAAANSUhEUgAAAX8AAAD7CAYAAACCEpQdAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n",
"text": [
"<matplotlib.figure.Figure at 0x1158b1e10>"
]
}
],
"prompt_number": 392
},
{
"cell_type": "code",
"collapsed": false,
"input": [
"loan_2.info()"
],
"language": "python",
"metadata": {},
"outputs": [
{
"output_type": "stream",
"stream": "stdout",
"text": [
"<class 'pandas.core.frame.DataFrame'>\n",
"Int64Index: 54419 entries, 3 to 188126\n",
"Data columns (total 9 columns):\n",
"funded_amnt 54415 non-null float64\n",
"emp_length 54415 non-null object\n",
"annual_inc 54415 non-null float64\n",
"loan_status 54415 non-null object\n",
"home_ownership 54415 non-null object\n",
"addr_state 54415 non-null object\n",
"tax_liens 54415 non-null float64\n",
"grade 54415 non-null object\n",
"loan_status_clean 54419 non-null int64\n",
"dtypes: float64(3), int64(1), object(5)"
]
}
],
"prompt_number": 393
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"###Cleaning Employment Length"
]
},
{
"cell_type": "code",
"collapsed": false,
"input": [
"loan_2['emp_length_clean'] = loan_2.emp_length.str.replace('+','')\n",
"loan_2['emp_length_clean'] = loan_2.emp_length_clean.str.replace('<','')\n",
"loan_2['emp_length_clean'] = loan_2.emp_length_clean.str.replace('years','')\n",
"loan_2['emp_length_clean'] = loan_2.emp_length_clean.str.replace('year','')\n",
"loan_2['emp_length_clean'] = loan_2.emp_length_clean.str.replace('n/a','0')"
],
"language": "python",
"metadata": {},
"outputs": [],
"prompt_number": 394
},
{
"cell_type": "code",
"collapsed": false,
"input": [
"loan_2.emp_length_clean.unique()"
],
"language": "python",
"metadata": {},
"outputs": [
{
"metadata": {},
"output_type": "pyout",
"prompt_number": 395,
"text": [
"array(['10 ', '2 ', '5 ', '1 ', '9 ', ' 1 ', '8 ', '0', '7 ', '4 ', '3 ',\n",
" '6 ', nan], dtype=object)"
]
}
],
"prompt_number": 395
},
{
"cell_type": "code",
"collapsed": false,
"input": [
"loan_2['emp_length_clean'] = loan_2.emp_length_clean.map(float)"
],
"language": "python",
"metadata": {},
"outputs": [],
"prompt_number": 396
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"###Cleaning Grade_Clean"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"Here, we'll be adding a value to letter grade assigned to individual loans. \"A\", the highest rating, will receive the value 7. \"G\", the lowest rating, will receive the value 1.\n",
"\n",
"Using the map function, we assign the values:"
]
},
{
"cell_type": "code",
"collapsed": false,
"input": [
"loan_2['grade_clean'] = loan_2['grade'].map({'A':7,'B':6,'C':5,'D':4,'E':3,'F':2,'G':1})"
],
"language": "python",
"metadata": {},
"outputs": [],
"prompt_number": 397
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"###Filling in mean values for NaN values."
]
},
{
"cell_type": "code",
"collapsed": false,
"input": [
"funded_amnt = loan_2.funded_amnt\n",
"mean_funded_amnt = loan_2[loan_2.funded_amnt.notnull()].funded_amnt.mean()\n",
"loan_2.funded_amnt.fillna(mean_funded_amnt, inplace=True)\n",
"\n",
"annual_inc = loan_2.annual_inc\n",
"mean_annual_inc = loan_2[loan_2.annual_inc.notnull()].annual_inc.mean()\n",
"loan_2.annual_inc.fillna(mean_annual_inc, inplace=True)\n",
"\n",
"emp_length = loan_2.emp_length_clean\n",
"mean_emp_length_clean = loan_2[loan_2.emp_length_clean.notnull()].emp_length_clean.mean()\n",
"loan_2.emp_length_clean.fillna(mean_emp_length_clean, inplace=True)\n",
"\n",
"grade = loan_2.grade\n",
"mean_grade_clean = loan_2[loan_2.grade.notnull()].grade_clean.mean()\n",
"loan_2.grade_clean.fillna(mean_grade_clean, inplace=True)"
],
"language": "python",
"metadata": {},
"outputs": [],
"prompt_number": 398
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"##Describing Data using Logistic Regression"
]
},
{
"cell_type": "code",
"collapsed": false,
"input": [
"import statsmodels.api as sm\n",
"from sklearn import linear_model, datasets\n",
"from sklearn.cross_validation import train_test_split"
],
"language": "python",
"metadata": {},
"outputs": [],
"prompt_number": 399
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"###First Logistic Regression (Employment Length & Grade of the Loan)\n",
"Predicting whether a loan will be paid off using Emplyoment Length and Grade of the Loan."
]
},
{
"cell_type": "code",
"collapsed": false,
"input": [
"X_Variables = ['emp_length_clean', 'grade_clean']\n",
"X = loan_2[X_Variables]"
],
"language": "python",
"metadata": {},
"outputs": [],
"prompt_number": 400
},
{
"cell_type": "code",
"collapsed": false,
"input": [
"X = X.values"
],
"language": "python",
"metadata": {},
"outputs": [],
"prompt_number": 401
},
{
"cell_type": "code",
"collapsed": false,
"input": [
"y = loan_2['loan_status_clean'].values"
],
"language": "python",
"metadata": {},
"outputs": [],
"prompt_number": 402
},
{
"cell_type": "code",
"collapsed": false,
"input": [],
"language": "python",
"metadata": {},
"outputs": [],
"prompt_number": 402
},
{
"cell_type": "code",
"collapsed": false,
"input": [
"clf = linear_model.LogisticRegression()"
],
"language": "python",
"metadata": {},
"outputs": [],
"prompt_number": 403
},
{
"cell_type": "code",
"collapsed": false,
"input": [
"model = clf.fit(X,y)"
],
"language": "python",
"metadata": {},
"outputs": [],
"prompt_number": 404
},
{
"cell_type": "code",
"collapsed": false,
"input": [
"model.score(X, y)"
],
"language": "python",
"metadata": {},
"outputs": [
{
"metadata": {},
"output_type": "pyout",
"prompt_number": 405,
"text": [
"0.77840460133409284"
]
}
],
"prompt_number": 405
},
{
"cell_type": "code",
"collapsed": false,
"input": [
"pd.DataFrame(zip(X_Variables,model.coef_.T))"
],
"language": "python",
"metadata": {},
"outputs": [
{
"html": [
"<div style=\"max-height:1000px;max-width:1500px;overflow:auto;\">\n",
"<table border=\"1\" class=\"dataframe\">\n",
" <thead>\n",
" <tr style=\"text-align: right;\">\n",
" <th></th>\n",
" <th>0</th>\n",
" <th>1</th>\n",
" </tr>\n",
" </thead>\n",
" <tbody>\n",
" <tr>\n",
" <th>0</th>\n",
" <td> emp_length_clean</td>\n",
" <td> [0.0159282068556]</td>\n",
" </tr>\n",
" <tr>\n",
" <th>1</th>\n",
" <td> grade_clean</td>\n",
" <td> [0.31113456483]</td>\n",
" </tr>\n",
" </tbody>\n",
"</table>\n",
"</div>"
],
"metadata": {},
"output_type": "pyout",
"prompt_number": 406,
"text": [
" 0 1\n",
"0 emp_length_clean [0.0159282068556]\n",
"1 grade_clean [0.31113456483]"
]
}
],
"prompt_number": 406
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"Above we have our coefficients:\n",
"\n",
"0.0159 for the lenght of employment\n",
"\n",
"0.3111 for the grade a loan received. "
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"Lets take a look at the grade a loan receives. For every additional increase in the grade \"G\" to \"F\" or in our case \"1\" to \"2\" the chance of the loan being paid off increases by .3111"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"Makes intuitive sense right? Why else would Lending Tree give a high grade to a loan that they think is faulty and as the grade for a loan increases so does the chance of the loan being paid off in this case by a coefficient of .3111"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"Alright. Well what about the years that someone has been employed? That certainly could be used as a predictor. In this case, it's not the best predictor. For every each additional year that someone is employed the chance of that person paying back their loan increases only by 0.0159 "
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"###Second Logistic Regression (Funded Amount & Annual Income)\n",
"Predicting whether a loan will be paid off using Funded Ammount and Annual Income. "
]
},
{
"cell_type": "code",
"collapsed": false,
"input": [
"X_Variables_2 = ['funded_amnt', 'annual_inc']\n",
"X_2 = loan_2[X_Variables_2]"
],
"language": "python",
"metadata": {},
"outputs": [],
"prompt_number": 407
},
{
"cell_type": "code",
"collapsed": false,
"input": [
"X_2 = X_2.values"
],
"language": "python",
"metadata": {},
"outputs": [],
"prompt_number": 408
},
{
"cell_type": "code",
"collapsed": false,
"input": [
"y_2 = loan_2['loan_status_clean'].values"
],
"language": "python",
"metadata": {},
"outputs": [],
"prompt_number": 409
},
{
"cell_type": "code",
"collapsed": false,
"input": [
"model_2 = clf.fit(X_2,y_2)"
],
"language": "python",
"metadata": {},
"outputs": [],
"prompt_number": 410
},
{
"cell_type": "code",
"collapsed": false,
"input": [
"model_2.score(X_2, y)"
],
"language": "python",
"metadata": {},
"outputs": [
{
"metadata": {},
"output_type": "pyout",
"prompt_number": 411,
"text": [
"0.77943365368713136"
]
}
],
"prompt_number": 411
},
{
"cell_type": "code",
"collapsed": false,
"input": [
"pd.DataFrame(zip(X_Variables_2,model_2.coef_.T))"
],
"language": "python",
"metadata": {},
"outputs": [
{
"html": [
"<div style=\"max-height:1000px;max-width:1500px;overflow:auto;\">\n",
"<table border=\"1\" class=\"dataframe\">\n",
" <thead>\n",
" <tr style=\"text-align: right;\">\n",
" <th></th>\n",
" <th>0</th>\n",
" <th>1</th>\n",
" </tr>\n",
" </thead>\n",
" <tbody>\n",
" <tr>\n",
" <th>0</th>\n",
" <td> funded_amnt</td>\n",
" <td> [-1.49216178225e-05]</td>\n",
" </tr>\n",
" <tr>\n",
" <th>1</th>\n",
" <td> annual_inc</td>\n",
" <td> [2.02376397324e-05]</td>\n",
" </tr>\n",
" </tbody>\n",
"</table>\n",
"</div>"
],
"metadata": {},
"output_type": "pyout",
"prompt_number": 412,
"text": [
" 0 1\n",
"0 funded_amnt [-1.49216178225e-05]\n",
"1 annual_inc [2.02376397324e-05]"
]
}
],
"prompt_number": 412
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"The reason for such low coefficients, for funded amount and annual income, is that the numbers are in thousands, granted they're in $, and loan status is binary ranging from 0 to 1. \n",
"\n",
"Let's look at the amount funded. As the amount funded increases by $10,000 the chance of it getting paid back decreases by -0.149 = (10,000 x -0.0000149).\n",
"\n",
"Similar, as annual income increases so does the chance of the loan being paid off. Intunitive, right? This is understandable and supported by the positive coefficient 0.0000202. In other words if my annual income increases by $10,000 so does the chance of me paying back the loan by 0.202 (10,000 x 0.0000202)"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"###Third Logistic Regression (Home Ownership)"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"Predicting whether a loan will be paid off given the individuals home ownership status. "
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"We currently have a column \"home_ownership\" with five unique values: \"Rent\", \"Mortgage\", \"Own\", \"None\", \"Other\"."
]
},
{
"cell_type": "code",
"collapsed": false,
"input": [
"loan_2.home_ownership.unique().tolist()"
],
"language": "python",
"metadata": {},
"outputs": [
{
"metadata": {},
"output_type": "pyout",
"prompt_number": 413,
"text": [
"['RENT', 'MORTGAGE', 'OWN', 'NONE', 'OTHER', nan]"
]
}
],
"prompt_number": 413
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"Before running this unique list, using logistic regression, against loan status, we have to create individual columns for each value, referred to as dummy variables. \n",
"\n",
"Each column will have a True or a False value associated with the individual loan that has either \"Rent\", \"Mortgage\", \"Own\", \"None\" or \"Other\"."
]
},
{
"cell_type": "code",
"collapsed": false,
"input": [
"home_ownership = pd.get_dummies(loan_2.home_ownership)\n",
"loan_2 = loan_2.join(home_ownership)"
],
"language": "python",
"metadata": {},
"outputs": [],
"prompt_number": 414
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"Below are our dummy variables."
]
},
{
"cell_type": "code",
"collapsed": false,
"input": [
"loan_2.head()"
],
"language": "python",
"metadata": {},
"outputs": [
{
"html": [
"<div style=\"max-height:1000px;max-width:1500px;overflow:auto;\">\n",
"<table border=\"1\" class=\"dataframe\">\n",
" <thead>\n",
" <tr style=\"text-align: right;\">\n",
" <th></th>\n",
" <th>funded_amnt</th>\n",
" <th>emp_length</th>\n",
" <th>annual_inc</th>\n",
" <th>loan_status</th>\n",
" <th>home_ownership</th>\n",
" <th>addr_state</th>\n",
" <th>tax_liens</th>\n",
" <th>grade</th>\n",
" <th>loan_status_clean</th>\n",
" <th>emp_length_clean</th>\n",
" <th>grade_clean</th>\n",
" <th>MORTGAGE</th>\n",
" <th>NONE</th>\n",
" <th>OTHER</th>\n",
" <th>OWN</th>\n",
" <th>RENT</th>\n",
" </tr>\n",
" </thead>\n",
" <tbody>\n",
" <tr>\n",
" <th>3 </th>\n",
" <td> 15000</td>\n",
" <td> 10+ years</td>\n",
" <td> 98000</td>\n",
" <td> Fully Paid</td>\n",
" <td> RENT</td>\n",
" <td> NY</td>\n",
" <td> 0</td>\n",
" <td> C</td>\n",
" <td> 1</td>\n",
" <td> 10</td>\n",
" <td> 5</td>\n",
" <td> 0</td>\n",
" <td> 0</td>\n",
" <td> 0</td>\n",
" <td> 0</td>\n",
" <td> 1</td>\n",
" </tr>\n",
" <tr>\n",
" <th>12</th>\n",
" <td> 3000</td>\n",
" <td> 10+ years</td>\n",
" <td> 25000</td>\n",
" <td> Fully Paid</td>\n",
" <td> RENT</td>\n",
" <td> FL</td>\n",
" <td> 0</td>\n",
" <td> B</td>\n",
" <td> 1</td>\n",
" <td> 10</td>\n",
" <td> 6</td>\n",
" <td> 0</td>\n",
" <td> 0</td>\n",
" <td> 0</td>\n",
" <td> 0</td>\n",
" <td> 1</td>\n",
" </tr>\n",
" <tr>\n",
" <th>15</th>\n",
" <td> 4800</td>\n",
" <td> 2 years</td>\n",
" <td> 39600</td>\n",
" <td> Fully Paid</td>\n",
" <td> MORTGAGE</td>\n",
" <td> TX</td>\n",
" <td> 0</td>\n",
" <td> B</td>\n",
" <td> 1</td>\n",
" <td> 2</td>\n",
" <td> 6</td>\n",
" <td> 1</td>\n",
" <td> 0</td>\n",
" <td> 0</td>\n",
" <td> 0</td>\n",
" <td> 0</td>\n",
" </tr>\n",
" <tr>\n",
" <th>18</th>\n",
" <td> 28000</td>\n",
" <td> 5 years</td>\n",
" <td> 325000</td>\n",
" <td> Fully Paid</td>\n",
" <td> MORTGAGE</td>\n",
" <td> CA</td>\n",
" <td> 0</td>\n",
" <td> A</td>\n",
" <td> 1</td>\n",
" <td> 5</td>\n",
" <td> 7</td>\n",
" <td> 1</td>\n",
" <td> 0</td>\n",
" <td> 0</td>\n",
" <td> 0</td>\n",
" <td> 0</td>\n",
" </tr>\n",
" <tr>\n",
" <th>22</th>\n",
" <td> 6000</td>\n",
" <td> 1 year</td>\n",
" <td> 70000</td>\n",
" <td> Fully Paid</td>\n",
" <td> MORTGAGE</td>\n",
" <td> NC</td>\n",
" <td> 0</td>\n",
" <td> B</td>\n",
" <td> 1</td>\n",
" <td> 1</td>\n",
" <td> 6</td>\n",
" <td> 1</td>\n",
" <td> 0</td>\n",
" <td> 0</td>\n",
" <td> 0</td>\n",
" <td> 0</td>\n",
" </tr>\n",
" </tbody>\n",
"</table>\n",
"</div>"
],
"metadata": {},
"output_type": "pyout",
"prompt_number": 415,
"text": [
" funded_amnt emp_length annual_inc loan_status home_ownership addr_state \\\n",
"3 15000 10+ years 98000 Fully Paid RENT NY \n",
"12 3000 10+ years 25000 Fully Paid RENT FL \n",
"15 4800 2 years 39600 Fully Paid MORTGAGE TX \n",
"18 28000 5 years 325000 Fully Paid MORTGAGE CA \n",
"22 6000 1 year 70000 Fully Paid MORTGAGE NC \n",
"\n",
" tax_liens grade loan_status_clean emp_length_clean grade_clean \\\n",
"3 0 C 1 10 5 \n",
"12 0 B 1 10 6 \n",
"15 0 B 1 2 6 \n",
"18 0 A 1 5 7 \n",
"22 0 B 1 1 6 \n",
"\n",
" MORTGAGE NONE OTHER OWN RENT \n",
"3 0 0 0 0 1 \n",
"12 0 0 0 0 1 \n",
"15 1 0 0 0 0 \n",
"18 1 0 0 0 0 \n",
"22 1 0 0 0 0 "
]
}
],
"prompt_number": 415
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"Let's run the logistic regression."
]
},
{
"cell_type": "code",
"collapsed": false,
"input": [
"X_Variables_3 = ['RENT', 'MORTGAGE', 'OWN', 'NONE', 'OTHER']\n",
"X_3 = loan_2[X_Variables_3]"
],
"language": "python",
"metadata": {},
"outputs": [],
"prompt_number": 416
},
{
"cell_type": "code",
"collapsed": false,
"input": [
"X_3 = X_3.values"
],
"language": "python",
"metadata": {},
"outputs": [],
"prompt_number": 417
},
{
"cell_type": "code",
"collapsed": false,
"input": [
"y_3 = loan_2['loan_status_clean'].values"
],
"language": "python",
"metadata": {},
"outputs": [],
"prompt_number": 418
},
{
"cell_type": "code",
"collapsed": false,
"input": [
"model_3 = clf.fit(X_3,y_3)"
],
"language": "python",
"metadata": {},
"outputs": [],
"prompt_number": 419
},
{
"cell_type": "code",
"collapsed": false,
"input": [
"model_3.score(X_3,y_3)"
],
"language": "python",
"metadata": {},
"outputs": [
{
"metadata": {},
"output_type": "pyout",
"prompt_number": 420,
"text": [
"0.77943365368713136"
]
}
],
"prompt_number": 420
},
{
"cell_type": "code",
"collapsed": false,
"input": [
"pd.DataFrame(zip(X_Variables_3, model_3.coef_.T))"
],
"language": "python",
"metadata": {},
"outputs": [
{
"html": [
"<div style=\"max-height:1000px;max-width:1500px;overflow:auto;\">\n",
"<table border=\"1\" class=\"dataframe\">\n",
" <thead>\n",
" <tr style=\"text-align: right;\">\n",
" <th></th>\n",
" <th>0</th>\n",
" <th>1</th>\n",
" </tr>\n",
" </thead>\n",
" <tbody>\n",
" <tr>\n",
" <th>0</th>\n",
" <td> RENT</td>\n",
" <td> [0.0310973077512]</td>\n",
" </tr>\n",
" <tr>\n",
" <th>1</th>\n",
" <td> MORTGAGE</td>\n",
" <td> [0.372819013643]</td>\n",
" </tr>\n",
" <tr>\n",
" <th>2</th>\n",
" <td> OWN</td>\n",
" <td> [0.135355977702]</td>\n",
" </tr>\n",
" <tr>\n",
" <th>3</th>\n",
" <td> NONE</td>\n",
" <td> [0.290465054095]</td>\n",
" </tr>\n",
" <tr>\n",
" <th>4</th>\n",
" <td> OTHER</td>\n",
" <td> [-0.794573804391]</td>\n",
" </tr>\n",
" </tbody>\n",
"</table>\n",
"</div>"
],
"metadata": {},
"output_type": "pyout",
"prompt_number": 421,
"text": [
" 0 1\n",
"0 RENT [0.0310973077512]\n",
"1 MORTGAGE [0.372819013643]\n",
"2 OWN [0.135355977702]\n",
"3 NONE [0.290465054095]\n",
"4 OTHER [-0.794573804391]"
]
}
],
"prompt_number": 421
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"My understanding for someone putting \"OTHER\" on their home ownership application is that either they did not want to reveal their home ownership situation, they didn't care or even bothered to fill it in on their application or in reality they just did't and don't know. Regardless, it seems that if someone puts \"OTHER\" on their home ownership application and gets funded, then there's a very good chance of that individual not paying back their loan. "
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"###Fourth Logistic Regression (Years Employed)"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"Using our column of years employed, we clreate dummies so that we could easily run the logistic regression. We're trying to see which length of employment is best predictive of someone paying back their loan."
]
},
{
"cell_type": "code",
"collapsed": false,
"input": [
"emp_dummies = pd.get_dummies(loan_2.emp_length)\n",
"loan_2 = loan_2.join(emp_dummies)"
],
"language": "python",
"metadata": {},
"outputs": [],
"prompt_number": 422
},
{
"cell_type": "code",
"collapsed": false,
"input": [
"X_Variables_4 = ['< 1 year','1 year','2 years','3 years','4 years','5 years','6 years','7 years','8 years','9 years','10+ years']\n",
"X_4 = loan_2[X_Variables_4]"
],
"language": "python",
"metadata": {},
"outputs": [],
"prompt_number": 423
},
{
"cell_type": "code",
"collapsed": false,
"input": [
"y_4 = loan_2['loan_status_clean'].values"
],
"language": "python",
"metadata": {},
"outputs": [],
"prompt_number": 424
},
{
"cell_type": "code",
"collapsed": false,
"input": [
"model_4 = clf.fit(X_4,y_4)"
],
"language": "python",
"metadata": {},
"outputs": [],
"prompt_number": 425
},
{
"cell_type": "code",
"collapsed": false,
"input": [
"model_4.score(X_4, y_4)"
],
"language": "python",
"metadata": {},
"outputs": [
{
"metadata": {},
"output_type": "pyout",
"prompt_number": 426,
"text": [
"0.77943365368713136"
]
}
],
"prompt_number": 426
},
{
"cell_type": "code",
"collapsed": false,
"input": [
"pd.DataFrame(zip(X_Variables_4,model_4.coef_.T))"
],
"language": "python",
"metadata": {},
"outputs": [
{
"html": [
"<div style=\"max-height:1000px;max-width:1500px;overflow:auto;\">\n",
"<table border=\"1\" class=\"dataframe\">\n",
" <thead>\n",
" <tr style=\"text-align: right;\">\n",
" <th></th>\n",
" <th>0</th>\n",
" <th>1</th>\n",
" </tr>\n",
" </thead>\n",
" <tbody>\n",
" <tr>\n",
" <th>0 </th>\n",
" <td> &lt; 1 year</td>\n",
" <td> [0.389953979006]</td>\n",
" </tr>\n",
" <tr>\n",
" <th>1 </th>\n",
" <td> 1 year</td>\n",
" <td> [0.462208894269]</td>\n",
" </tr>\n",
" <tr>\n",
" <th>2 </th>\n",
" <td> 2 years</td>\n",
" <td> [0.512777795606]</td>\n",
" </tr>\n",
" <tr>\n",
" <th>3 </th>\n",
" <td> 3 years</td>\n",
" <td> [0.468445970599]</td>\n",
" </tr>\n",
" <tr>\n",
" <th>4 </th>\n",
" <td> 4 years</td>\n",
" <td> [0.393890046838]</td>\n",
" </tr>\n",
" <tr>\n",
" <th>5 </th>\n",
" <td> 5 years</td>\n",
" <td> [0.444452229474]</td>\n",
" </tr>\n",
" <tr>\n",
" <th>6 </th>\n",
" <td> 6 years</td>\n",
" <td> [0.403099349594]</td>\n",
" </tr>\n",
" <tr>\n",
" <th>7 </th>\n",
" <td> 7 years</td>\n",
" <td> [0.42611970081]</td>\n",
" </tr>\n",
" <tr>\n",
" <th>8 </th>\n",
" <td> 8 years</td>\n",
" <td> [0.420555034092]</td>\n",
" </tr>\n",
" <tr>\n",
" <th>9 </th>\n",
" <td> 9 years</td>\n",
" <td> [0.412132834046]</td>\n",
" </tr>\n",
" <tr>\n",
" <th>10</th>\n",
" <td> 10+ years</td>\n",
" <td> [0.496784122381]</td>\n",
" </tr>\n",
" </tbody>\n",
"</table>\n",
"</div>"
],
"metadata": {},
"output_type": "pyout",
"prompt_number": 427,
"text": [
" 0 1\n",
"0 < 1 year [0.389953979006]\n",
"1 1 year [0.462208894269]\n",
"2 2 years [0.512777795606]\n",
"3 3 years [0.468445970599]\n",
"4 4 years [0.393890046838]\n",
"5 5 years [0.444452229474]\n",
"6 6 years [0.403099349594]\n",
"7 7 years [0.42611970081]\n",
"8 8 years [0.420555034092]\n",
"9 9 years [0.412132834046]\n",
"10 10+ years [0.496784122381]"
]
}
],
"prompt_number": 427
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"There doesn't seem to be too much variance between the generated coefficients of the years employed. It looks like, so long as the person is employed they will be paying back their loan. \n",
"\n",
"However, it holds true, that if someone is unemployed or has less than a year of employment then they'll have a lower chance of repaying their loan. "
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"####For some graphs and charts refer to my earlier post on this same dataset. Below I'll be going over some mapping using cartodb which is a web mapping tool. "
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"##Describing Data using CartoDB"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"Below we're generating a csv file, with select columns, so that we could pull it into cartodb. Cartodb is a web mapping tool. "
]
},
{
"cell_type": "code",
"collapsed": false,
"input": [
"loan_cartodb2 = loan_2[['addr_state','funded_amnt','emp_length_clean','annual_inc','grade_clean','loan_status_clean']]\n",
"#loan_cartodb2.to_csv('/Users/olehdubno/Desktop/loan_cartodb2.csv', index=False)"
],
"language": "python",
"metadata": {},
"outputs": [],
"prompt_number": 428
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"Here we use www.cartodb.com. A very intuitive and friendly way of generating maps. \n",
"\n",
"Before mapping our data, cartodb uses an intelligent way, in this instance, in converting State acronyms into latitude and longitude.\n",
"\n",
"After selecting the features we want to play with, cartodb generates a map and ways to share it. One of the ways is using IFrame. IFrame uses HTML to embed content from one source into another."
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"###Mapping Paid and Unpaid Loans"
]
},
{
"cell_type": "code",
"collapsed": false,
"input": [
"from IPython.display import HTML\n",
"HTML(\"<iframe width='100%' height='520' frameborder='0' src='http://shortyskater456.cartodb.com/viz/40d16f7e-6d3e-11e4-a898-0e9d821ea90d/embed_map' allowfullscreen webkitallowfullscreen mozallowfullscreen oallowfullscreen msallowfullscreen></iframe>\")"
],
"language": "python",
"metadata": {},
"outputs": [
{
"html": [
"<iframe width='100%' height='520' frameborder='0' src='http://shortyskater456.cartodb.com/viz/40d16f7e-6d3e-11e4-a898-0e9d821ea90d/embed_map' allowfullscreen webkitallowfullscreen mozallowfullscreen oallowfullscreen msallowfullscreen></iframe>"
],
"metadata": {},
"output_type": "pyout",
"prompt_number": 429,
"text": [
"<IPython.core.display.HTML at 0x1179f86d0>"
]
}
],
"prompt_number": 429
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"The above map is referred to as the choropleth map, \"a thematic map in which areas are shade patterned in proportion to the measurement of the statistical variable being displayed.\" (wikipedia)\n",
"\n",
"As the intensity of the color increases (gets closer to 1), on average the majority of the people residing in that state have paid of their loan.\n",
"\n",
"The number near the point references the amount of loans given in that state.\n",
"\n",
"By the looks of the map I wouldn't give loans out to Oregon, Wisconsin, Nevada, Tennessee, Virginia, Indianapolis, maybe a few others.\n",
"\n",
"Of course this an average of individual loans, per state, discounting specific regions of the state, and is not the best estimate for whether a funded individual in that state is likely to repay their loan.\n",
"\n",
"However, maybe the other features could help determine which state is less liklier to pay off a loan. "
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"###Mapping The Amount Funded"
]
},
{
"cell_type": "code",
"collapsed": false,
"input": [
"HTML(\"<iframe width='100%' height='520' frameborder='0' src='http://olehdubno.cartodb.com/viz/0dce85a4-6d10-11e4-98f3-0e9d821ea90d/embed_map' allowfullscreen webkitallowfullscreen mozallowfullscreen oallowfullscreen msallowfullscreen></iframe>\")"
],
"language": "python",
"metadata": {},
"outputs": [
{
"html": [
"<iframe width='100%' height='520' frameborder='0' src='http://olehdubno.cartodb.com/viz/0dce85a4-6d10-11e4-98f3-0e9d821ea90d/embed_map' allowfullscreen webkitallowfullscreen mozallowfullscreen oallowfullscreen msallowfullscreen></iframe>"
],
"metadata": {},
"output_type": "pyout",
"prompt_number": 430,
"text": [
"<IPython.core.display.HTML at 0x1179f8f50>"
]
}
],
"prompt_number": 430
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"Reviewing the map we could see that Oregon, Montana and Mississippi, on average, have taken out higher loans (closer to $35,000).\n",
"\n",
"According to the \"Paid vs Unpaid\" map, Oregon is not only taking out the highest loans, it's also not paying them back.\n",
"\n",
"On average, indivduals receiving a loan in Oregon are much more liklier to default on their loan as they are also liklier to receive bigger loans. "
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"###Mapping Annual Income"
]
},
{
"cell_type": "code",
"collapsed": false,
"input": [
"HTML(\"<iframe width='100%' height='520' frameborder='0' src='http://olehdubno.cartodb.com/viz/c2c9b8a4-6ba6-11e4-aadc-0e4fddd5de28/embed_map' allowfullscreen webkitallowfullscreen mozallowfullscreen oallowfullscreen msallowfullscreen></iframe>\")"
],
"language": "python",
"metadata": {},
"outputs": [
{
"html": [
"<iframe width='100%' height='520' frameborder='0' src='http://olehdubno.cartodb.com/viz/c2c9b8a4-6ba6-11e4-aadc-0e4fddd5de28/embed_map' allowfullscreen webkitallowfullscreen mozallowfullscreen oallowfullscreen msallowfullscreen></iframe>"
],
"metadata": {},
"output_type": "pyout",
"prompt_number": 431,
"text": [
"<IPython.core.display.HTML at 0x1184de1d0>"
]
}
],
"prompt_number": 431
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"There's several outliers in the data that have been removed, using cartodb, in terms of annual income. \n",
"\n",
"Before removing the outliers, the income ranges from $33,504.72 to $7,241,778. Which is an obsene amount. I limit it to $500,000.00\n",
"\n",
"Interestingly, Oregon is the state with the highest income, lowest payback rate and on average the state that takes out the highest loans. "
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"###Mapping The Grade Given To Individual Loans"
]
},
{
"cell_type": "code",
"collapsed": false,
"input": [
"HTML(\"<iframe width='100%' height='520' frameborder='0' src='http://olehdubno.cartodb.com/viz/57bfeb6c-6ba8-11e4-a74d-0e4fddd5de28/embed_map' allowfullscreen webkitallowfullscreen mozallowfullscreen oallowfullscreen msallowfullscreen></iframe>\")"
],
"language": "python",
"metadata": {},
"outputs": [
{
"html": [
"<iframe width='100%' height='520' frameborder='0' src='http://olehdubno.cartodb.com/viz/57bfeb6c-6ba8-11e4-a74d-0e4fddd5de28/embed_map' allowfullscreen webkitallowfullscreen mozallowfullscreen oallowfullscreen msallowfullscreen></iframe>"
],
"metadata": {},
"output_type": "pyout",
"prompt_number": 432,
"text": [
"<IPython.core.display.HTML at 0x11767f050>"
]
}
],
"prompt_number": 432
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"Keeping on track with Oregon, a state I'm not too familiar with, it happens to be a state with a fairly good rating for loans according to the data, not really. At least for the loans given out by Lending Club.\n",
"\n",
"I could understand why Lending Club, on average, would give a pretty good grade to loans in Oregon. The average population there has some of the highest income. We could see that by looking at the income map presented before."
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"##Decision Tree"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"####Train Test Split\n",
"Here, we're using the train test split function from sklearn to split up the features into train and test values. Our test size will be 25% of our actual data."
]
},
{
"cell_type": "code",
"collapsed": false,
"input": [
"X_Variables = ['emp_length_clean', 'grade_clean']\n",
"X = loan_2[X_Variables]\n",
"\n",
"X = X.values\n",
"y = loan_2['loan_status_clean'].values"
],
"language": "python",
"metadata": {},
"outputs": [],
"prompt_number": 481
},
{
"cell_type": "code",
"collapsed": false,
"input": [
"from sklearn.cross_validation import train_test_split\n",
"\n",
"X_train, X_test, Y_train, Y_test = train_test_split(X,y,test_size=0.25)"
],
"language": "python",
"metadata": {},
"outputs": [],
"prompt_number": 482
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"clf.fit cretes a classifier object, which is called \"clf', and then this new \"fitted\" object, \"clf\", can do things like score and predict."
]
},
{
"cell_type": "code",
"collapsed": false,
"input": [
"clf = GaussianNB()"
],
"language": "python",
"metadata": {},
"outputs": [],
"prompt_number": 483
},
{
"cell_type": "code",
"collapsed": false,
"input": [
"clf.fit(X_train,Y_train)"
],
"language": "python",
"metadata": {},
"outputs": [
{
"metadata": {},
"output_type": "pyout",
"prompt_number": 484,
"text": [
"GaussianNB()"
]
}
],
"prompt_number": 484
},
{
"cell_type": "code",
"collapsed": false,
"input": [
"clf.score(X_train,Y_train)"
],
"language": "python",
"metadata": {},
"outputs": [
{
"metadata": {},
"output_type": "pyout",
"prompt_number": 485,
"text": [
"0.78019797128436319"
]
}
],
"prompt_number": 485
},
{
"cell_type": "code",
"collapsed": false,
"input": [
"clf.score(X_test,Y_test)"
],
"language": "python",
"metadata": {},
"outputs": [
{
"metadata": {},
"output_type": "pyout",
"prompt_number": 486,
"text": [
"0.77126056596839399"
]
}
],
"prompt_number": 486
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"###Time for Decision Tree"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"After "
]
},
{
"cell_type": "code",
"collapsed": false,
"input": [
"from sklearn.naive_bayes import GaussianNB\n",
"clf = GaussianNB()\n",
"clf.fit(X_train,Y_train)"
],
"language": "python",
"metadata": {},
"outputs": [
{
"metadata": {},
"output_type": "pyout",
"prompt_number": 487,
"text": [
"GaussianNB()"
]
}
],
"prompt_number": 487
},
{
"cell_type": "code",
"collapsed": false,
"input": [
"from sklearn import metrics\n",
"def measure_performance(X,y,clf, show_accuracy=True, show_classification_report=True, show_confusion_matrix=True):\n",
" y_pred=clf.predict(X) \n",
" if show_accuracy:\n",
" print \"Accuracy:{0:.3f}\".format(metrics.accuracy_score(y,y_pred)),\"\\n\"\n",
"\n",
" if show_classification_report:\n",
" print \"Classification report\"\n",
" print metrics.classification_report(y,y_pred),\"\\n\"\n",
" \n",
" if show_confusion_matrix:\n",
" print \"Confusion matrix\"\n",
" print metrics.confusion_matrix(y,y_pred),\"\\n\"\n",
" \n",
"measure_performance(X_train,Y_train,clf, show_classification_report=True, show_confusion_matrix=True)"
],
"language": "python",
"metadata": {},
"outputs": [
{
"output_type": "stream",
"stream": "stdout",
"text": [
"Accuracy:0.780 \n",
"\n",
"Classification report\n",
" precision recall f1-score support\n",
"\n",
" 0 0.42 0.01 0.03 8919\n",
" 1 0.78 0.99 0.88 31895\n",
"\n",
"avg / total 0.70 0.78 0.69 40814\n",
"\n",
"\n",
"Confusion matrix\n",
"[[ 127 8792]\n",
" [ 179 31716]] \n",
"\n"
]
}
],
"prompt_number": 488
},
{
"cell_type": "code",
"collapsed": false,
"input": [
"from IPython.display import Image\n",
"Image(filename='/Users/olehdubno/Desktop/confusion_matrix.png')"
],
"language": "python",
"metadata": {},
"outputs": [
{
"metadata": {},
"output_type": "pyout",
"png": "iVBORw0KGgoAAAANSUhEUgAAAgsAAAB6CAIAAAC+3IfUAAAYWGlDQ1BJQ0MgUHJvZmlsZQAAWAmt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\niywigLWQRfBZ1lgLLBRZDLBa+P8HaKI5Ccm59AAAAABJRU5ErkJggg==\n",
"prompt_number": 440,
"text": [
"<IPython.core.display.Image at 0x1044737d0>"
]
}
],
"prompt_number": 440
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"Confusion Matrix allows for more detailed analysis than mere proportion of correct guesses.\n",
"\n",
"For instance 177 loans from paid loans were incorrecly predicted as unpaid. \n",
"\n",
"Based on the entries in the confusion matrix, the total number of correct predictions made by the model is (177 loans + 31,594 loans) and the total number of incorrect predictions is (177 loans + 8,920 loans).\n",
"\n",
"The confusion matrix provides the information needed to determine how well a classification model performs. The perforamnce metric, accuracy, summarizes this information with a single number .777 \n",
"\n",
"Accuracy takes the total number of correct predictions and divides it by the total number of all predictions made. "
]
},
{
"cell_type": "code",
"collapsed": false,
"input": [
"predictions = [p[1] for p in clf.predict_proba(X_train)]\n",
"fpr_p, tpr_p, thresholds_p = metrics.roc_curve(Y_train,predictions)\n",
"\n",
"fig = plt.figure()\n",
"fig.set_figwidth(10)\n",
"fig.suptitle('AUC for Decision Tree Classifier Predicting Loans Paid')\n",
"\n",
"ax1 = plt.subplot(1, 2, 1)\n",
"ax1.set_xlabel('false positive rate')\n",
"ax1.set_ylabel('true positive rate')\n",
"ax1.plot(fpr_p, tpr_p)\n",
"\n",
"fpr, tpr, thresholds = metrics.roc_curve(Y_train,clf.predict(X_train))\n",
"ax2 = plt.subplot(1, 2, 2)\n",
"ax2.set_xlabel('false positive rate')\n",
"ax2.set_ylabel('true positive rate')\n",
"ax2.plot(fpr, tpr)\n",
"\n",
"\n",
"print \"False-positive rate:\", fpr\n",
"print \"True-positive rate: \", tpr\n",
"print \"Thresholds: \", thresholds\n",
"\n",
"print fig"
],
"language": "python",
"metadata": {},
"outputs": [
{
"output_type": "stream",
"stream": "stdout",
"text": [
"False-positive rate: [ 0. 0.98576074 1. ]\n",
"True-positive rate: [ 0. 0.99438784 1. ]\n",
"Thresholds: [2 1 0]\n",
"Figure(800x320)\n"
]
},
{
"metadata": {},
"output_type": "display_data",
"png": "iVBORw0KGgoAAAANSUhEUgAAAmUAAAEhCAYAAAA+rtfAAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nYII=\n",
"text": [
"<matplotlib.figure.Figure at 0x117a4a510>"
]
}
],
"prompt_number": 489
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"#Naive Bayes Classifier"
]
},
{
"cell_type": "code",
"collapsed": false,
"input": [
"from sklearn.naive_bayes import GaussianNB\n",
"from sklearn.metrics import accuracy_score"
],
"language": "python",
"metadata": {},
"outputs": [],
"prompt_number": 441
},
{
"cell_type": "code",
"collapsed": false,
"input": [
"clf = GaussianNB()"
],
"language": "python",
"metadata": {},
"outputs": [],
"prompt_number": 442
},
{
"cell_type": "code",
"collapsed": false,
"input": [
"clf.fit(X_train,Y_train)"
],
"language": "python",
"metadata": {},
"outputs": [
{
"metadata": {},
"output_type": "pyout",
"prompt_number": 443,
"text": [
"GaussianNB()"
]
}
],
"prompt_number": 443
},
{
"cell_type": "code",
"collapsed": false,
"input": [
"pred = clf.predict(X_test)"
],
"language": "python",
"metadata": {},
"outputs": [],
"prompt_number": 444
},
{
"cell_type": "code",
"collapsed": false,
"input": [
"accuracy_score(pred, Y_test)"
],
"language": "python",
"metadata": {},
"outputs": [
{
"metadata": {},
"output_type": "pyout",
"prompt_number": 445,
"text": [
"0.77677324513046675"
]
}
],
"prompt_number": 445
},
{
"cell_type": "code",
"collapsed": false,
"input": [],
"language": "python",
"metadata": {},
"outputs": [],
"prompt_number": 445
},
{
"cell_type": "code",
"collapsed": false,
"input": [],
"language": "python",
"metadata": {},
"outputs": [],
"prompt_number": 445
},
{
"cell_type": "code",
"collapsed": false,
"input": [],
"language": "python",
"metadata": {},
"outputs": [],
"prompt_number": 445
},
{
"cell_type": "code",
"collapsed": false,
"input": [],
"language": "python",
"metadata": {},
"outputs": [],
"prompt_number": 445
},
{
"cell_type": "code",
"collapsed": false,
"input": [],
"language": "python",
"metadata": {},
"outputs": [],
"prompt_number": 445
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"one notebook is the everything notebook (all the code) (developer notebook)\n",
"\n",
"second is the subset of images and explanatory text and sum code if helpful that could illustrate the presentation in conversanalist way. (presentation notebook) (client-facing notebook)\n",
"you could use code snippets in the client facing notebook\n",
"\n",
"\n",
"\n"
]
},
{
"cell_type": "code",
"collapsed": false,
"input": [
"loan_2.describe()"
],
"language": "python",
"metadata": {},
"outputs": [
{
"html": [
"<div style=\"max-height:1000px;max-width:1500px;overflow:auto;\">\n",
"<table border=\"1\" class=\"dataframe\">\n",
" <thead>\n",
" <tr style=\"text-align: right;\">\n",
" <th></th>\n",
" <th>funded_amnt</th>\n",
" <th>annual_inc</th>\n",
" <th>tax_liens</th>\n",
" <th>loan_status_clean</th>\n",
" <th>emp_length_clean</th>\n",
" <th>grade_clean</th>\n",
" <th>MORTGAGE</th>\n",
" <th>NONE</th>\n",
" <th>OTHER</th>\n",
" <th>OWN</th>\n",
" <th>...</th>\n",
" <th>2 years</th>\n",
" <th>3 years</th>\n",
" <th>4 years</th>\n",
" <th>5 years</th>\n",
" <th>6 years</th>\n",
" <th>7 years</th>\n",
" <th>8 years</th>\n",
" <th>9 years</th>\n",
" <th>&lt; 1 year</th>\n",
" <th>n/a</th>\n",
" </tr>\n",
" </thead>\n",
" <tbody>\n",
" <tr>\n",
" <th>count</th>\n",
" <td> 54419.000000</td>\n",
" <td> 54419.000000</td>\n",
" <td> 54415.000000</td>\n",
" <td> 54419.000000</td>\n",
" <td> 54419.000000</td>\n",
" <td> 54419.000000</td>\n",
" <td> 54419.000000</td>\n",
" <td> 54419.000000</td>\n",
" <td> 54419.000000</td>\n",
" <td> 54419.000000</td>\n",
" <td>...</td>\n",
" <td> 54419.000000</td>\n",
" <td> 54419.000000</td>\n",
" <td> 54419.000000</td>\n",
" <td> 54419.000000</td>\n",
" <td> 54419.000000</td>\n",
" <td> 54419.000000</td>\n",
" <td> 54419.000000</td>\n",
" <td> 54419.000000</td>\n",
" <td> 54419.000000</td>\n",
" <td> 54419.000000</td>\n",
" </tr>\n",
" <tr>\n",
" <th>mean</th>\n",
" <td> 13924.277313</td>\n",
" <td> 71833.822999</td>\n",
" <td> 0.009483</td>\n",
" <td> 0.779434</td>\n",
" <td> 5.734044</td>\n",
" <td> 5.134853</td>\n",
" <td> 0.496738</td>\n",
" <td> 0.000294</td>\n",
" <td> 0.000221</td>\n",
" <td> 0.079641</td>\n",
" <td>...</td>\n",
" <td> 0.092137</td>\n",
" <td> 0.075654</td>\n",
" <td> 0.064941</td>\n",
" <td> 0.081699</td>\n",
" <td> 0.069057</td>\n",
" <td> 0.060310</td>\n",
" <td> 0.045499</td>\n",
" <td> 0.037524</td>\n",
" <td> 0.075984</td>\n",
" <td> 0.035080</td>\n",
" </tr>\n",
" <tr>\n",
" <th>std</th>\n",
" <td> 8094.307473</td>\n",
" <td> 61003.894957</td>\n",
" <td> 0.142897</td>\n",
" <td> 0.414632</td>\n",
" <td> 3.541428</td>\n",
" <td> 1.336344</td>\n",
" <td> 0.499994</td>\n",
" <td> 0.017145</td>\n",
" <td> 0.014848</td>\n",
" <td> 0.270740</td>\n",
" <td>...</td>\n",
" <td> 0.289222</td>\n",
" <td> 0.264446</td>\n",
" <td> 0.246423</td>\n",
" <td> 0.273909</td>\n",
" <td> 0.253553</td>\n",
" <td> 0.238062</td>\n",
" <td> 0.208397</td>\n",
" <td> 0.190043</td>\n",
" <td> 0.264976</td>\n",
" <td> 0.183983</td>\n",
" </tr>\n",
" <tr>\n",
" <th>min</th>\n",
" <td> 1000.000000</td>\n",
" <td> 4800.000000</td>\n",
" <td> 0.000000</td>\n",
" <td> 0.000000</td>\n",
" <td> 0.000000</td>\n",
" <td> 1.000000</td>\n",
" <td> 0.000000</td>\n",
" <td> 0.000000</td>\n",
" <td> 0.000000</td>\n",
" <td> 0.000000</td>\n",
" <td>...</td>\n",
" <td> 0.000000</td>\n",
" <td> 0.000000</td>\n",
" <td> 0.000000</td>\n",
" <td> 0.000000</td>\n",
" <td> 0.000000</td>\n",
" <td> 0.000000</td>\n",
" <td> 0.000000</td>\n",
" <td> 0.000000</td>\n",
" <td> 0.000000</td>\n",
" <td> 0.000000</td>\n",
" </tr>\n",
" <tr>\n",
" <th>25%</th>\n",
" <td> 8000.000000</td>\n",
" <td> 45000.000000</td>\n",
" <td> 0.000000</td>\n",
" <td> 1.000000</td>\n",
" <td> 2.000000</td>\n",
" <td> 4.000000</td>\n",
" <td> 0.000000</td>\n",
" <td> 0.000000</td>\n",
" <td> 0.000000</td>\n",
" <td> 0.000000</td>\n",
" <td>...</td>\n",
" <td> 0.000000</td>\n",
" <td> 0.000000</td>\n",
" <td> 0.000000</td>\n",
" <td> 0.000000</td>\n",
" <td> 0.000000</td>\n",
" <td> 0.000000</td>\n",
" <td> 0.000000</td>\n",
" <td> 0.000000</td>\n",
" <td> 0.000000</td>\n",
" <td> 0.000000</td>\n",
" </tr>\n",
" <tr>\n",
" <th>50%</th>\n",
" <td> 12000.000000</td>\n",
" <td> 62000.000000</td>\n",
" <td> 0.000000</td>\n",
" <td> 1.000000</td>\n",
" <td> 6.000000</td>\n",
" <td> 5.000000</td>\n",
" <td> 0.000000</td>\n",
" <td> 0.000000</td>\n",
" <td> 0.000000</td>\n",
" <td> 0.000000</td>\n",
" <td>...</td>\n",
" <td> 0.000000</td>\n",
" <td> 0.000000</td>\n",
" <td> 0.000000</td>\n",
" <td> 0.000000</td>\n",
" <td> 0.000000</td>\n",
" <td> 0.000000</td>\n",
" <td> 0.000000</td>\n",
" <td> 0.000000</td>\n",
" <td> 0.000000</td>\n",
" <td> 0.000000</td>\n",
" </tr>\n",
" <tr>\n",
" <th>75%</th>\n",
" <td> 19100.000000</td>\n",
" <td> 85000.000000</td>\n",
" <td> 0.000000</td>\n",
" <td> 1.000000</td>\n",
" <td> 10.000000</td>\n",
" <td> 6.000000</td>\n",
" <td> 1.000000</td>\n",
" <td> 0.000000</td>\n",
" <td> 0.000000</td>\n",
" <td> 0.000000</td>\n",
" <td>...</td>\n",
" <td> 0.000000</td>\n",
" <td> 0.000000</td>\n",
" <td> 0.000000</td>\n",
" <td> 0.000000</td>\n",
" <td> 0.000000</td>\n",
" <td> 0.000000</td>\n",
" <td> 0.000000</td>\n",
" <td> 0.000000</td>\n",
" <td> 0.000000</td>\n",
" <td> 0.000000</td>\n",
" </tr>\n",
" <tr>\n",
" <th>max</th>\n",
" <td> 35000.000000</td>\n",
" <td> 7141778.000000</td>\n",
" <td> 10.000000</td>\n",
" <td> 1.000000</td>\n",
" <td> 10.000000</td>\n",
" <td> 7.000000</td>\n",
" <td> 1.000000</td>\n",
" <td> 1.000000</td>\n",
" <td> 1.000000</td>\n",
" <td> 1.000000</td>\n",
" <td>...</td>\n",
" <td> 1.000000</td>\n",
" <td> 1.000000</td>\n",
" <td> 1.000000</td>\n",
" <td> 1.000000</td>\n",
" <td> 1.000000</td>\n",
" <td> 1.000000</td>\n",
" <td> 1.000000</td>\n",
" <td> 1.000000</td>\n",
" <td> 1.000000</td>\n",
" <td> 1.000000</td>\n",
" </tr>\n",
" </tbody>\n",
"</table>\n",
"<p>8 rows \u00d7 23 columns</p>\n",
"</div>"
],
"metadata": {},
"output_type": "pyout",
"prompt_number": 446,
"text": [
" funded_amnt annual_inc tax_liens loan_status_clean \\\n",
"count 54419.000000 54419.000000 54415.000000 54419.000000 \n",
"mean 13924.277313 71833.822999 0.009483 0.779434 \n",
"std 8094.307473 61003.894957 0.142897 0.414632 \n",
"min 1000.000000 4800.000000 0.000000 0.000000 \n",
"25% 8000.000000 45000.000000 0.000000 1.000000 \n",
"50% 12000.000000 62000.000000 0.000000 1.000000 \n",
"75% 19100.000000 85000.000000 0.000000 1.000000 \n",
"max 35000.000000 7141778.000000 10.000000 1.000000 \n",
"\n",
" emp_length_clean grade_clean MORTGAGE NONE \\\n",
"count 54419.000000 54419.000000 54419.000000 54419.000000 \n",
"mean 5.734044 5.134853 0.496738 0.000294 \n",
"std 3.541428 1.336344 0.499994 0.017145 \n",
"min 0.000000 1.000000 0.000000 0.000000 \n",
"25% 2.000000 4.000000 0.000000 0.000000 \n",
"50% 6.000000 5.000000 0.000000 0.000000 \n",
"75% 10.000000 6.000000 1.000000 0.000000 \n",
"max 10.000000 7.000000 1.000000 1.000000 \n",
"\n",
" OTHER OWN ... 2 years 3 years \\\n",
"count 54419.000000 54419.000000 ... 54419.000000 54419.000000 \n",
"mean 0.000221 0.079641 ... 0.092137 0.075654 \n",
"std 0.014848 0.270740 ... 0.289222 0.264446 \n",
"min 0.000000 0.000000 ... 0.000000 0.000000 \n",
"25% 0.000000 0.000000 ... 0.000000 0.000000 \n",
"50% 0.000000 0.000000 ... 0.000000 0.000000 \n",
"75% 0.000000 0.000000 ... 0.000000 0.000000 \n",
"max 1.000000 1.000000 ... 1.000000 1.000000 \n",
"\n",
" 4 years 5 years 6 years 7 years 8 years \\\n",
"count 54419.000000 54419.000000 54419.000000 54419.000000 54419.000000 \n",
"mean 0.064941 0.081699 0.069057 0.060310 0.045499 \n",
"std 0.246423 0.273909 0.253553 0.238062 0.208397 \n",
"min 0.000000 0.000000 0.000000 0.000000 0.000000 \n",
"25% 0.000000 0.000000 0.000000 0.000000 0.000000 \n",
"50% 0.000000 0.000000 0.000000 0.000000 0.000000 \n",
"75% 0.000000 0.000000 0.000000 0.000000 0.000000 \n",
"max 1.000000 1.000000 1.000000 1.000000 1.000000 \n",
"\n",
" 9 years < 1 year n/a \n",
"count 54419.000000 54419.000000 54419.000000 \n",
"mean 0.037524 0.075984 0.035080 \n",
"std 0.190043 0.264976 0.183983 \n",
"min 0.000000 0.000000 0.000000 \n",
"25% 0.000000 0.000000 0.000000 \n",
"50% 0.000000 0.000000 0.000000 \n",
"75% 0.000000 0.000000 0.000000 \n",
"max 1.000000 1.000000 1.000000 \n",
"\n",
"[8 rows x 23 columns]"
]
}
],
"prompt_number": 446
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"#OLS on funded amount and annual income"
]
},
{
"cell_type": "code",
"collapsed": false,
"input": [
"loan_limit_by_inc = loan_2[loan_2['annual_inc']<50000]"
],
"language": "python",
"metadata": {},
"outputs": [],
"prompt_number": 455
},
{
"cell_type": "code",
"collapsed": false,
"input": [],
"language": "python",
"metadata": {},
"outputs": []
},
{
"cell_type": "code",
"collapsed": false,
"input": [
"loan_limit_by_inc.info()"
],
"language": "python",
"metadata": {},
"outputs": [
{
"output_type": "stream",
"stream": "stdout",
"text": [
"<class 'pandas.core.frame.DataFrame'>\n",
"Int64Index: 16683 entries, 12 to 188121\n",
"Data columns (total 28 columns):\n",
"funded_amnt 16683 non-null float64\n",
"emp_length 16683 non-null object\n",
"annual_inc 16683 non-null float64\n",
"loan_status 16683 non-null object\n",
"home_ownership 16683 non-null object\n",
"addr_state 16683 non-null object\n",
"tax_liens 16683 non-null float64\n",
"grade 16683 non-null object\n",
"loan_status_clean 16683 non-null int64\n",
"emp_length_clean 16683 non-null float64\n",
"grade_clean 16683 non-null float64\n",
"MORTGAGE 16683 non-null float64\n",
"NONE 16683 non-null float64\n",
"OTHER 16683 non-null float64\n",
"OWN 16683 non-null float64\n",
"RENT 16683 non-null float64\n",
"1 year 16683 non-null float64\n",
"10+ years 16683 non-null float64\n",
"2 years 16683 non-null float64\n",
"3 years 16683 non-null float64\n",
"4 years 16683 non-null float64\n",
"5 years 16683 non-null float64\n",
"6 years 16683 non-null float64\n",
"7 years 16683 non-null float64\n",
"8 years 16683 non-null float64\n",
"9 years 16683 non-null float64\n",
"< 1 year 16683 non-null float64\n",
"n/a 16683 non-null float64\n",
"dtypes: float64(22), int64(1), object(5)"
]
}
],
"prompt_number": 456
},
{
"cell_type": "code",
"collapsed": false,
"input": [
"import statsmodels.formula.api as smf\n",
"# OLS, or ordinary least squares, takes a y (dependent variable) and X (independent variables) (formula = y ~ X)\n",
"# Below, we copy the data frame and remove the na variables, and create a single variable linear model\n",
"# to return a test statistic and p-value, to see how strong of a relationship bodyweight and brainweight have.\n",
"\n",
"loan_limit_by_inc['log_annual_inc'] = np.log(loan_limit_by_inc['annual_inc'])\n",
"loan_limit_by_inc['log_funded_amnt'] = np.log(loan_limit_by_inc['funded_amnt'])\n",
"\n",
"fig, axes = plt.subplots(nrows=1,ncols=2)\n",
"\n",
"axes[0].plot(loan_limit_by_inc.annual_inc, loan_limit_by_inc.funded_amnt, 'go')\n",
"\n",
"model = smf.ols(formula='funded_amnt ~ annual_inc', data=loan_limit_by_inc)\n",
"results = model.fit()\n",
"print 'NORMAL FIT SUMMARY'\n",
"print(results.summary())\n",
"print\n",
"\n",
"axes[1].plot(loan_limit_by_inc.log_annual_inc, loan_limit_by_inc.log_funded_amnt, 'mo')\n",
"\n",
"log_model = smf.ols(formula='log_funded_amnt ~ log_annual_inc', data=loan_limit_by_inc)\n",
"log_results = log_model.fit()\n",
"print 'LOG-LOG FIT SUMMARY'\n",
"print(log_results.summary())\n",
"\n",
"print fig"
],
"language": "python",
"metadata": {},
"outputs": [
{
"output_type": "stream",
"stream": "stdout",
"text": [
"NORMAL FIT SUMMARY\n",
" OLS Regression Results \n",
"==============================================================================\n",
"Dep. Variable: funded_amnt R-squared: 0.201\n",
"Model: OLS Adj. R-squared: 0.201\n",
"Method: Least Squares F-statistic: 1.346e+04\n",
"Date: Tue, 18 Nov 2014 Prob (F-statistic): 0.00\n",
"Time: 20:04:00 Log-Likelihood: -5.5040e+05\n",
"No. Observations: 53491 AIC: 1.101e+06\n",
"Df Residuals: 53489 BIC: 1.101e+06\n",
"Df Model: 1 \n",
"==============================================================================\n",
" coef std err t P>|t| [95.0% Conf. Int.]\n",
"------------------------------------------------------------------------------\n",
"Intercept 6142.0323 72.528 84.685 0.000 5999.877 6284.187\n",
"annual_inc 0.1122 0.001 116.035 0.000 0.110 0.114\n",
"==============================================================================\n",
"Omnibus: 1305.245 Durbin-Watson: 1.979\n",
"Prob(Omnibus): 0.000 Jarque-Bera (JB): 1403.149\n",
"Skew: 0.397 Prob(JB): 2.04e-305\n",
"Kurtosis: 2.990 Cond. No. 1.77e+05\n",
"==============================================================================\n",
"\n",
"Warnings:\n",
"[1] The condition number is large, 1.77e+05. This might indicate that there are\n",
"strong multicollinearity or other numerical problems.\n",
"\n",
"LOG-LOG FIT SUMMARY"
]
},
{
"output_type": "stream",
"stream": "stdout",
"text": [
"\n",
" OLS Regression Results \n",
"==============================================================================\n",
"Dep. Variable: log_funded_amnt R-squared: 0.201\n",
"Model: OLS Adj. R-squared: 0.201\n",
"Method: Least Squares F-statistic: 1.349e+04\n",
"Date: Tue, 18 Nov 2014 Prob (F-statistic): 0.00\n",
"Time: 20:04:00 Log-Likelihood: -48154.\n",
"No. Observations: 53491 AIC: 9.631e+04\n",
"Df Residuals: 53489 BIC: 9.633e+04\n",
"Df Model: 1 \n",
"==================================================================================\n",
" coef std err t P>|t| [95.0% Conf. Int.]\n",
"----------------------------------------------------------------------------------\n",
"Intercept 2.2661 0.061 37.187 0.000 2.147 2.385\n",
"log_annual_inc 0.6418 0.006 116.164 0.000 0.631 0.653\n",
"==============================================================================\n",
"Omnibus: 7184.471 Durbin-Watson: 1.986\n",
"Prob(Omnibus): 0.000 Jarque-Bera (JB): 11032.520\n",
"Skew: -0.965 Prob(JB): 0.00\n",
"Kurtosis: 4.108 Cond. No. 263.\n",
"=============================================================================="
]
},
{
"output_type": "stream",
"stream": "stdout",
"text": [
"\n",
"Figure(480x320)\n"
]
},
{
"metadata": {},
"output_type": "display_data",
"png": "iVBORw0KGgoAAAANSUhEUgAAAYkAAAEACAYAAABGYoqtAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nAElFTkSuQmCC\n",
"text": [
"<matplotlib.figure.Figure at 0x1158aa5d0>"
]
}
],
"prompt_number": 368
},
{
"cell_type": "code",
"collapsed": false,
"input": [
"#multilinear regression OLS\n",
"\n",
"import pandas as pd\n",
"import numpy as np\n",
"import statsmodels.api as sm\n",
"\n",
"X = loan_limit_by_inc[['annual_inc','emp_length_clean','grade_clean']]\n",
"y = loan_limit_by_inc['funded_amnt']\n"
],
"language": "python",
"metadata": {},
"outputs": [],
"prompt_number": 460
},
{
"cell_type": "code",
"collapsed": false,
"input": [
"X = sm.add_constant(X)\n",
"est = sm.OLS(y,X).fit()"
],
"language": "python",
"metadata": {},
"outputs": [],
"prompt_number": 461
},
{
"cell_type": "code",
"collapsed": false,
"input": [
"est.summary()"
],
"language": "python",
"metadata": {},
"outputs": [
{
"html": [
"<table class=\"simpletable\">\n",
"<caption>OLS Regression Results</caption>\n",
"<tr>\n",
" <th>Dep. Variable:</th> <td>funded_amnt</td> <th> R-squared: </th> <td> 0.212</td> \n",
"</tr>\n",
"<tr>\n",
" <th>Model:</th> <td>OLS</td> <th> Adj. R-squared: </th> <td> 0.212</td> \n",
"</tr>\n",
"<tr>\n",
" <th>Method:</th> <td>Least Squares</td> <th> F-statistic: </th> <td> 1495.</td> \n",
"</tr>\n",
"<tr>\n",
" <th>Date:</th> <td>Tue, 18 Nov 2014</td> <th> Prob (F-statistic):</th> <td> 0.00</td> \n",
"</tr>\n",
"<tr>\n",
" <th>Time:</th> <td>21:42:26</td> <th> Log-Likelihood: </th> <td>-1.6328e+05</td>\n",
"</tr>\n",
"<tr>\n",
" <th>No. Observations:</th> <td> 16683</td> <th> AIC: </th> <td>3.266e+05</td> \n",
"</tr>\n",
"<tr>\n",
" <th>Df Residuals:</th> <td> 16679</td> <th> BIC: </th> <td>3.266e+05</td> \n",
"</tr>\n",
"<tr>\n",
" <th>Df Model:</th> <td> 3</td> <th> </th> <td> </td> \n",
"</tr>\n",
"</table>\n",
"<table class=\"simpletable\">\n",
"<tr>\n",
" <td></td> <th>coef</th> <th>std err</th> <th>t</th> <th>P>|t|</th> <th>[95.0% Conf. Int.]</th> \n",
"</tr>\n",
"<tr>\n",
" <th>const</th> <td> 2194.4996</td> <td> 212.309</td> <td> 10.336</td> <td> 0.000</td> <td> 1778.351 2610.648</td>\n",
"</tr>\n",
"<tr>\n",
" <th>annual_inc</th> <td> 0.2503</td> <td> 0.004</td> <td> 60.796</td> <td> 0.000</td> <td> 0.242 0.258</td>\n",
"</tr>\n",
"<tr>\n",
" <th>emp_length_clean</th> <td> 78.6857</td> <td> 9.810</td> <td> 8.021</td> <td> 0.000</td> <td> 59.456 97.915</td>\n",
"</tr>\n",
"<tr>\n",
" <th>grade_clean</th> <td> -478.7410</td> <td> 27.957</td> <td> -17.124</td> <td> 0.000</td> <td> -533.541 -423.941</td>\n",
"</tr>\n",
"</table>\n",
"<table class=\"simpletable\">\n",
"<tr>\n",
" <th>Omnibus:</th> <td>197.206</td> <th> Durbin-Watson: </th> <td> 1.967</td>\n",
"</tr>\n",
"<tr>\n",
" <th>Prob(Omnibus):</th> <td> 0.000</td> <th> Jarque-Bera (JB): </th> <td> 137.111</td>\n",
"</tr>\n",
"<tr>\n",
" <th>Skew:</th> <td> 0.106</td> <th> Prob(JB): </th> <td>1.69e-30</td>\n",
"</tr>\n",
"<tr>\n",
" <th>Kurtosis:</th> <td> 2.609</td> <th> Cond. No. </th> <td>2.43e+05</td>\n",
"</tr>\n",
"</table>"
],
"metadata": {},
"output_type": "pyout",
"prompt_number": 462,
"text": [
"<class 'statsmodels.iolib.summary.Summary'>\n",
"\"\"\"\n",
" OLS Regression Results \n",
"==============================================================================\n",
"Dep. Variable: funded_amnt R-squared: 0.212\n",
"Model: OLS Adj. R-squared: 0.212\n",
"Method: Least Squares F-statistic: 1495.\n",
"Date: Tue, 18 Nov 2014 Prob (F-statistic): 0.00\n",
"Time: 21:42:26 Log-Likelihood: -1.6328e+05\n",
"No. Observations: 16683 AIC: 3.266e+05\n",
"Df Residuals: 16679 BIC: 3.266e+05\n",
"Df Model: 3 \n",
"====================================================================================\n",
" coef std err t P>|t| [95.0% Conf. Int.]\n",
"------------------------------------------------------------------------------------\n",
"const 2194.4996 212.309 10.336 0.000 1778.351 2610.648\n",
"annual_inc 0.2503 0.004 60.796 0.000 0.242 0.258\n",
"emp_length_clean 78.6857 9.810 8.021 0.000 59.456 97.915\n",
"grade_clean -478.7410 27.957 -17.124 0.000 -533.541 -423.941\n",
"==============================================================================\n",
"Omnibus: 197.206 Durbin-Watson: 1.967\n",
"Prob(Omnibus): 0.000 Jarque-Bera (JB): 137.111\n",
"Skew: 0.106 Prob(JB): 1.69e-30\n",
"Kurtosis: 2.609 Cond. No. 2.43e+05\n",
"==============================================================================\n",
"\n",
"Warnings:\n",
"[1] The condition number is large, 2.43e+05. This might indicate that there are\n",
"strong multicollinearity or other numerical problems.\n",
"\"\"\""
]
}
],
"prompt_number": 462
},
{
"cell_type": "code",
"collapsed": false,
"input": [],
"language": "python",
"metadata": {},
"outputs": []
}
],
"metadata": {}
}
]
}
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment