More
Сhoose
Role: Risk Analyst
Tools: Python, SQL, Matplotlib, Seaborn
Category: Financial Risk
Case Study Overview
Analyzing loan applicant profiles and historical defaults to help a credit provider minimize non-performing loans (NPLs) and standardize credit score rules.
Problem Statement
Identify applicant segments showing high tendencies of defaults, ensuring safer loan approvals.
Analytical Focus
Structuring relational SQL schemas, filtering client segments, and modeling credit metrics using Python datasets.
The Analytical Approach
By querying transactional credit history databases, I isolated data for default status indicators. I calculated interest ratios, debt-to-income limits, and repayment delays.
Python was utilized to execute cohort analysis, plotting debt distribution curves and determining default probabilities across customer income groups.
Results & Business Value
Key Metric Improvement
8% Reduction in Non-Performing Loans
- Isolated debt-to-income threshold ratios above which defaults rise exponentially.
- Enabled credit approval officers to filter applications using a simplified risk matrix.
- Improved financial auditing turnaround times through automated reporting scripts.