TimDuu / Loan-Risk-Analysis-Lending-Club

I do the loan risk analysis for Lending Club. It is a typical end to end data science problem. I Analyze the loan risk through predicting interest rate and final loan status.

Geek Repo:Geek Repo

Github PK Tool:Github PK Tool

Loan-Risk-Analysis-Lending-Club

The .RMD files show the code I used to do analysis.
The .PDF files show the result of analysis.

Reading sequence would be

  1. LoanRiskAnalysis_EDA.pdf or LoanRiskAnalysis_EDA.Rmd
  2. LoanRiskAnalysis_InterestRate.pdf or LoanRiskAnalysis_InterestRate.Rmd
  3. LoanRiskAnalysis_LoanStatus.pdf or LoanRiskAnalysis_LoanStatus.Rmd

Summary of data analysis
Part one: I explored three questions in LoanRiskAnalysis_EDA part.
■ What factor will influence the interest rate?
■ What are the distribution of loan status?
■ What are the purpose of applying a loan with LendingClub?

Part two: I engineered feature and built linear regression in LoanRiskAnalysis_InterestRate part.
■ Performed correlation analysis for numeric features and hypothesis test (unpaired t-test, ANOVA) for category features.
■ Built multi-variable linear regression model for interest rate prediction.
■ Selected high impact features through Lasso regularization.
■ Evaluated the performance of interest rate model through adjusted coefficient of determination.

Part three: I engineered feature and built logistic regression in LoanRiskAnalysis_LoanStatus part.
■ Performed correlation analysis for numeric features and hypothesis test (unpaired t-test, ANOVA) for category features.
■ Built logistic regression model for interest rate prediction.
■ Selected high impact features through Lasso regularization.
■ Evaluated the performance of loan status model by ROC/AUC.

About

I do the loan risk analysis for Lending Club. It is a typical end to end data science problem. I Analyze the loan risk through predicting interest rate and final loan status.