Ahomka/credit-portfolio-monitor

Python-based credit portfolio monitoring system for banks. Tracks NPL ratios, loan classifications, collateral coverage, provision adequacy, and remediation performance. Includes an Early Warning Scoring model and automated monitoring report generation. Aligned with Bank of Ghana prudential guidelines. # 🏦 Ghana Bank - Credit Portfolio Monitoring System > A Python-based credit monitoring analytics system replicating the core workflows of a bank Credit Monitoring Officer: NPL tracking, portfolio classification, early warning scoring, provision analysis, and automated report generation. --- ## 📊 Dashboard Preview --- ## 🎯 What This Project Does This project simulates the analytical environment of a bank credit monitoring function and directly addresses the KPIs used to evaluate a Credit Monitoring Officer: | JD KPI | Implementation | |---|---| | **NPL as % of Total Loans** | Computed and tracked by classification bucket | | **Quality & Number of Reviews / Portfolio Coverage** | Review completion rate calculated per account | | **Timely Delivery & Quality of Monitoring Reports** | Automated CSV monitoring reports generated on run | | **Planned vs. Achieved Improvement Process Initiative** | Remediation target vs. achieved tracking by class | | **Reduction in Impaired Assets / Bad Debt Portfolio** | Remediation value analysis with sector breakdown | | **Remediation Value to Staff Cost** | Proxy ratio computed against estimated staff cost base | --- ## 🔬 Analytical Modules ### 1. Portfolio Classification & NPL Ratio Tracking Classifies the full loan portfolio using the Bank of Ghana 5-tier system: - **Current** → 1% provision - **Watch** → 5% provision - **Substandard** → 25% provision - **Doubtful** → 50% provision - **Loss** → 100% provision Calculates NPL ratio, provision adequacy, and collateral coverage per class. ### 2. Sector & Branch Concentration Risk Identifies which sectors carry the highest NPL ratios and flags concentration risk. Outputs a ranked view of sector exposure for targeted remediation. ### 3. Early Warning Scoring (EWS) Model A rules-based scoring model (0-100) that flags high-risk accounts based on: - Days Past Due (DPD) - weighted 40 points - Collateral Coverage - weighted 30 points - Provision Rate - weighted 20 points - Review …