Uncovering Lending Patterns

In this project, I used Power BI to analyze a structured dataset of loan applications and visualize the key variables that affect loan decisions. The aim was to understand how financial institutions might assess risk beyond simple metrics like credit score.

Project Goals:

• Determine if higher credit scores lead to more approvals
• Analyze the impact of home ownership types on decisions
• Explore how bankruptcies and tax liens influence rejection patterns

Methodology:

• Data cleaning in Power BI

• Custom visuals (bar charts, pie charts, matrix tables)

• Slicers to filter across variables (credit score, ownership, etc.)

• Conditional formatting to highlight high-risk combinations

Key Insights:

Credit Score ≠ Guaranteed Approval

• Higher credit scores don’t ensure loan approval.
• Factors like income, current debts, and job stability often override score alone.

Home Ownership & Loan Status

• Mortgage holders are most common but don’t have the highest approval rate.
• Homeowners with no mortgage tend to get more approvals.
• Renters are the most denied - possibly due to less asset backing.

Bankruptcy & Tax Liens

• Surprisingly, applicants with 0 bankruptcies or tax liens don’t always have higher approval.
• Risk rises with more bankruptcies, but some 1-lien or 1-bankruptcy applicants fare better than expected.
• A combined analysis showed the most risky pair is 2 bankruptcies and 1 lien.

Data Fields Used:

Loan status, credit history, home ownership, income, debt levels, credit problems, bankruptcies, tax liens.

Who Can Use This:

• Bank risk analysts looking for real-world decision patterns
• Lending institutions evaluating non-linear risk variables
• Policy developers refining automated lending rules

Watch dashboard demo