An End-to-End Data Engineering & Business Intelligence Solution
In the financial sector, tracking Non-Performing Loans (NPL) isn't just about making charts—it is about risk mitigation and capital preservation.
This project transforms raw, disconnected banking data into a production-ready, cross-filtering analytical engine. Designed with enterprise-grade data architecture, it allows banking executives to instantly isolate bad debt ratios down to the individual branch level, shifting credit risk management from reactive to proactive.
Great executive dashboards are built on flawless backend data models.
To ensure optimal query performance and highly scalable DAX calculations, the backend was strictly modeled using a Star Schema architecture. This establishes a single source of truth and prevents the performance bottlenecks common in flat-file reporting.
- 🧱 Fact Table (
fact_loans): The quantitative core, tracking millions in loan amounts, outstanding balances, and risk provisions. - 🌐 Conformed Dimensions: Dedicated lookups for
dim_branches,dim_customers, and a robust time-intelligence calendar. - 📂 Semantic Layer (
_Risk Measures): A hollow, disconnected measure table housing all custom DAX logic (e.g., NPL Ratio, Total Portfolio Value) to keep the calculation engine organized and scalable.
The frontend reporting layer was designed with a high-contrast, minimalist aesthetic to draw immediate attention to critical risk outliers.
- Dynamic Cross-Filtering: Selecting a specific branch instantly recalculates all KPI cards and redraws the trend lines for that exact location.
- Time Intelligence: Year-over-year NPL tracking powered by the active calendar dimension relationship.
- Actionable Insights: Discovered and isolated a critical 10.88% NPL ratio at specific Abu Dhabi branches against a portfolio average of 7.67%.
Architected and engineered by Data With RISHAL
Transforming raw data into strategic business assets.