Objectives: This project investigates which borrower and loan attributes most strongly influence loan repayment behavior. The goal is to provide Prosper and similar lending platforms with actionable insights to enhance credit risk assessment and improve loan approval strategies.
Tools: Used Python (with pandas and NumPy) for data preprocessing, Matplotlib and Seaborn for data visualization, and Power BI for building an interactive dashboard.
Strategies and Techniques: Performed data cleaning, univariate and bivariate exploratory data analysis to uncover relationships between borrower characteristics and loan outcomes, and developed a visual dashboard to highlight loan default risks.
Results: Identified BorrowerAPR, StatedMonthlyIncome, and EmploymentStatus as strong predictors of default risk. Found that term length and recommendation count had minimal impact on repayment outcomes.
Conclusions: The analysis suggests focusing more on borrowers' income levels and employment status during loan evaluations. Adjusting interest rate thresholds accordingly could help minimize default risk and improve overall loan portfolio performance.