A Real-Time Disaster Intelligence and Evacuation Support Platform Using Geospatial Visualization, Live Event Synchronization, and Climate Risk Inference

Natural disasters create rapidly changing information environments in which citizens, volunteers, and administrators require a common operational picture. This paper presents a full-stack prototype for real-time disaster tracking and evacuation support. The platform integrates a React-based geospatial interface, a Python Flask and Flask-SocketIO backend, JSON persistence, role-aware authentication, browser-based live location, OpenWeather current-weather data, road-network routing, multilingual emergency guidance, SOS signaling, and PDF situation-report generation. Administrators can publish incident and response-resource information, while citizens can view hazards, locate resources, request evacuation routes, and transmit distress information. A transparent climate-risk heuristic combines rainfall, wind speed, and humidity into a bounded preliminary screening score. Functional evaluation was performed across authentication, incident and shelter workflows, climate analysis, live-location display, weather-data integration, administrator-to-user synchronization, routing, SOS, and reporting. The observed tests demonstrate integrated operation of these components; however, no numerical accuracy claim is made for geolocation, route estimation, or the heuristic because dedicated ground-truth and historical validation datasets were not part of the supplied evaluation. The principal contribution is an integrated, explainable emergency-coordination workflow rather than a new prediction algorithm.