Netra (नेत्र — "The Eye") is a full-scale AI-Driven Crime Analytics & Predictive Policing Command Platform for State Law Enforcement agencies.
| Layer | Technology |
|---|---|
| Frontend | React 18 + TypeScript + Vite |
| Styling | Tailwind CSS v4 |
| Charts | Recharts |
| Aurora FX | Canvas API + requestAnimationFrame |
| Backend | Python FastAPI |
| AI/ML | XGBoost + scikit-learn + spaCy (design spec) |
| Graph DB | Neo4j (design spec) |
| GIS | PostGIS + Mapbox GL (design spec) |
- 🌌 Aurora Gradient Hero — 60 FPS canvas WebGL-style interactive gradient with spring cursor glow
- 🔐 Multi-role RBAC — District SP, SHO, Analyst, Auditor, DGP, Public Citizen
- 🗺️ Geospatial Hotspot Map — Time slider, crime type filters, incident inspector
- 🕸️ Link & Network Analysis — Multi-hop entity graph (suspects, FIRs, vehicles, SIMs)
- 🧠 Explainable Predictive Risk — SHAP factor attribution transparency + ethics audit
- 🎛️ What-If Resource Simulator — Patrol allocation sandbox with live risk forecasting
- 🤖 AI Co-pilot — Natural language query engine (RAG + LLM query parsing)
- 📋 WORM Audit Trail — Cryptographic immutable audit log for oversight compliance
- 👁️ Public Transparency Portal — Anonymized 1km² aggregated safety heatmaps
This platform is designed with ethical AI first principles:
- All AI risk scores include SHAP-based human-readable explanation panels
- Human-in-the-loop review required before patrol dispatch
- Independent Oversight Board auditor role with read-only cryptographic log access
- Disparate Impact (DI) metric fairness auditing built into the model pipeline
© 2026 State Police Cyber Cell & Intelligence Analytics Wing