Disaster Response Coordination Platform

Backend-Heavy MERN Stack Project

---

Overview

A real-time coordination system built to manage disaster response using AI-driven insights, geospatial queries, and live updates.
It consolidates citizen reports, relief resources, and official data into a single, intelligent platform.

---

Motivation

During disasters, information is scattered across multiple platforms — social media, news, and official sites.
My motivation was to build a unified system that merges these streams, enhances them with AI, and helps responders act faster with reliable, location-based intelligence.

---

Problem It Solves

• Centralizes real-time data from citizens, authorities, and social media.
• Extracts and verifies locations or images using Google Gemini AI.
• Maps nearby resources with Supabase/Postgres geospatial queries.
• Broadcasts live updates via Socket.IO.
• Caches data intelligently to reduce redundant API calls.

---

Tech Stack

Category	Technology
Frontend	React.js
Backend	Node.js, Express.js
Database	Supabase (PostgreSQL + Geospatial)
AI Integration	Google Gemini API
Real-time	Socket.IO
Caching & Storage	Supabase JSONB
Maps/Geocoding	Google Maps API / OpenStreetMap
Deployment	Render / Fly.io / Railway (Backend), Vercel / Netlify (Frontend)

---

Features

• 🧠 AI-Powered Location Extraction: Extracts coordinates from text using Gemini.
• 🖼️ Image Verification: Validates authenticity through AI models.
• 🗺️ Geospatial Resource Mapping: Finds resources near disaster areas.
• ⚡ Realtime Updates: Broadcasts events using Socket.IO.
• 📰 Official Updates Scraper: Fetches verified government/NGO reports.
• 💾 Smart Caching: Optimizes response time with Supabase caching.
• 💬 Mock Social Feed: Simulates citizen updates for testing.

---

What I Learned

• Designing AI-integrated backend systems with multiple APIs.
• Writing geospatial SQL queries and Supabase RPC functions.
• Handling real-time communication efficiently with Socket.IO.
• Deploying backend-heavy apps using cloud-native tools.

---

What Makes It Stand Out

• AI + Geospatial + Real-time — all integrated in one system.
• Backend-first architecture optimized for reliability and scalability.
• Intelligent caching and indexing for speed and efficiency.
• Minimal frontend, but powerful backend workflows designed for real-world disaster coordination.

---

Quick Start

Backend

cd backend
cp .env.example .env   # Fill in environment variables
npm install
npm run dev
# Runs on http://localhost:5000

Frontend

cd frontend
npm install
npm start

Runs on http://localhost:3000

Environment Variables (Backend)

PORT=5000
SUPABASE_URL=
SUPABASE_ANON_KEY=
GEMINI_API_KEY=
GOOGLE_MAPS_API_KEY=

FRONTEND_URL=http://localhost:3000

Database Setup

ables: disasters, resources, reports, cache
Spatial columns: disasters.location, resources.location
Indexes: GIST (spatial), GIN (tags)
RPC: nearby_resources(lon, lat, radius_m, p_disaster_id)
SQL file: backend/database-setup.sql