segfault| Just wait on per page due to api connecting time| will be sorted in final deployment

- Self-healing digital twin of Chennai — 68,495 zones, 90,670 road and drainage corridors
- Real-time 6-factor physics simulation: flood, traffic, AQI, heat, rainfall, congestion
- Graph Attention Network (GAT) predicts flood and risk propagation across real city topology
- Bidirectional LSTM forecasts conditions at t+5 to t+25 minutes with confidence scores
- PPO reinforcement learning agent autonomously intervenes — drainage, rerouting, green corridors
- Adaptive Wardrop router distributes 100 users across 13 routes naturally — no central coordinator
- Causal attribution engine explains why zones are at risk using GNN attention weights
- Dual interface — citizen dashboard (public) and operator console (JWT authenticated)
- Live OpenWeatherMap integration — real Chennai weather feeds the simulation every 5 minutes
- Built on real OSMnx road graph — not synthetic data

Видео segfault| Just wait on per page due to api connecting time| will be sorted in final deployment канала Ayushman Yashaswi