Lending Club Case Study (Consumer Finance Company)

Consumer Finance Company is specialised in lending different types of loans to customers.
This case study aims to understand the driving factors behind loan defaulters, which can help in risk assessment of the company.

Table of Contents

• General Info
• Technologies Used
• Conclusions
• Recommendations
• Acknowledgements

General Information

• Background: Lending Club provides personal loans to individuals. The company wants to minimize losses due to loan defaulters by identifying key factors that lead to default.
• Business Problem: The goal is to analyze historical loan data to identify patterns and variables that are strong predictors of loan defaulters. This will help the company improve its credit risk assessment and lending policies.
• Dataset: The dataset (loan.csv) contains information about past loan applicants, including loan amount, term, interest rate, employment length, home ownership, annual income, loan status, and more. The data has been cleansed to remove irrelevant columns, standardize formats, and handle missing values.

Technologies Used

• Python 3.x
• pandas
• numpy
• matplotlib
• seaborn
• Jupyter Notebook

Conclusions

Key Findings

• Loan Grade and Sub-Grade:
  Default rates increase sharply for lower grades (especially F and G). Sub-grades G3, G5, and F5 show the highest default rates. Grade is a strong summary indicator of risk, as it incorporates multiple borrower and loan characteristics.

• Loan Purpose:
  Certain purposes are consistently associated with higher default rates, especially:

  • Small business
  • Renewable energy
  • Medical
  • House
  • Educational
    These purposes are even riskier when combined with large loan amounts or low grades.

• Loan Amount:
  Higher loan amounts are associated with higher default rates across most purposes. The risk is especially pronounced for small business, renewable energy, and medical loans at higher amounts.

• Interest Rate:
  Higher interest rates are correlated with higher default rates, as riskier borrowers are charged higher rates.

• Annual Income:
  Most defaults occur in the 30k–65k income range (due to volume), but lower income alone is not a strong predictor after controlling for other variables.

High Risk Combinations

Purpose	High Risk Grades	High Risk Loan Amounts
small business	D, E, F, G	Highest bins (largest loans)
renewable energy	F, G	Highest bins
medical	F, G	High bins
house	F, G	High bins
educational	F, G	High bins

Recommendations

• Lending policies should be more conservative for loans with:
  • Low grades
  • High loan amounts
  • High risk purposes (especially small business, renewable energy, and medical)
  • High terms and Interest Rates
• These combinations have the highest observed default rates and should be prioritized for stricter credit assessment and monitoring.

Acknowledgements

• Inspired by Lending Club’s open data and peer-to-peer lending research.
• Data dictionary and business context from Lending Club documentation.
• This project was based on MTech AI/ML Course Lending Case Study.