DATA 607 02 [15961] : Final Project

Debabrata Kabiraj
05/12/2019

Introduction

Lending Club (LC), a San Francisco-based fintech company, works to facilitate peer-to-peer loans through their online lending platform. Started in 2007, their website allows individuals to publicly post loan applications, which other users can then browse and choose to fund. Company estimates place aggregate loan totals at over $15.98 billion through December 2015, making Lending Club the largest online loan platform in the world.

Simply put, LC is a US peer-to-peer lending company. Where investors provide funding and borrowers return back the payments. Lending club selects and approves the borrowers using many parameters. It is a sort of EBay for loans.

Project Goal

This project is to predict the interest rate with various predictor variables. By performing this analysis we will know below information.

  1. What parameters will impact my interest rate? ie., Is loan interest % predictive of FICO credit score alone?

  2. Is loan funded amount are equal for different purpose of loan request? So the person can get loan in that particular loan type.

  3. It is always mentioned that living state plays a important role in interest rate. This hypothesis will be validated.

  4. There is a myth that home ownership will impact FICO scores. It will be validated via this dataset.

  5. Did lending club receive equal number of loans in each month?

Data Source

When we register as a lending club user, you will get access to the borrowers data from lending loan website Lending Club.

This dataset has borrowers details (personal info will be removed) It has the funded amount, interest rate, fico credit score and about 150 variables. Also the row count is around 115K for Q1 2019.

Data Collection

Download the Loan Statistics from Lending Club WebPage via LendingClub's API.

Sample Data Post Data CleanUp

Data Exploration

We know that grades have values of A, B, C, D, E, F, G where A represents the highest quality loan and G the lowest.

Grades

LOAN Issued By Grades

LOAN-ISSUED-GROUPED-BY-GRADES

Interest Rate by Grades

Interest Rate by Grade

LOAN Issued By FICO Scores

LOAN-ISSUED-GROUPED-BY-FICO-SCORES

FICO vs Loan Interest Rate

FICOvsIntRate

Home Ownership VS FICO Scores

Home Ownership VS FICO Scores

Interest Rate Histogram

Interest Rate Histogram

Annual Income VS Funded Amount

Annual Income VS Funded Amount

Installment vs Funded Amount

Installment vs Funded Amount

Purpose VS Funded Amount

Purpose VS Funded Amount

Loan Amount VS Purpose for Loan Status

Loan Amount VS Purpose for Loan Status

Loan Amount by Grade for each Loan Status

Loan Amount by Grade for each Loan Status

State VS Average Interest Rate

State Vs Average Interest Rate

Distribution of Interest Rates across US States

Distribution of Interest Rates across US States

Distribution of Loan Amounts across US States

Distribution of Interest Rates across US States

Conclusion

  • Dataset from Lending club give insights of interest rate from Bank which provides interesting information about the interest rate for each person.

  • The interest rate depends on the various factors like FICO score, Homeownership, Purpose of loan, Term length of loan, loan amount requested, Annual income, Employee length, Issue month, Previous bankruptcies and Debt to Income ratio.

  • If a person is wanting to get a good interest rate then he need to focus on above factors before applying for a lending club loan.

  • Future analysis can be performed by adding more variables to the model to determine at-risk loans or Charged-off loans. We can also calculate the return interest rate for an investor via Recommendation System with more variable data to train the system.

Q & A

QnA