MedicaidLens
An interactive dashboard that makes $1.09 Trillion in healthcare spending transparent, allowing investigators to visually explore ML-flagged fraud anomalies.
Tags
Technologies
Overview
MedicaidLens is an interactive web application designed to make $1.09 Trillion in public healthcare spending transparent, explorable, and accountable using machine learning.
It ingests vast amounts of Medicaid claims and provider data, joining it with machine learning risk scores (generated by MedicaidGuard) to detect and explain potential healthcare fraud, waste, and abuse.
🚀 Features
- National Overview Dashboard: High-level statistical summaries and geographic risk distribution.
- Provider Drill-down: Detailed specific provider profiles showing ML confidence scores, statistical anomaly flags, and interactive SHAP value explanations for why a score was assigned.
- Interactive Geospatial Hotspots: State-by-state heatmaps identifying regional spending anomalies.
- Watchlist & Search: Filter and export tabular lists of high-risk providers for further investigation.
- Procedure Code Benchmarks: Analyze national spending distributions by specific medical service codes.
🛠️ Data Pipeline Integration
MedicaidLens is designed to consume the output of the MedicaidGuard machine learning pipeline. The high-performance FastAPI back-end reads directly from optimized Parquet tables using DuckDB, allowing lightning-fast query resolution to serve the dynamic React front-end dashboard.