kenenisabeyan/sfews-ethiopia

Smart Flood Early Warning System (SFEWS) using IoT and Machine Learning for the Awash River Basin. # 🌊 SFEWS — Smart Flood Early Warning System > An AI-powered, real-time flood monitoring and forecasting platform for Ethiopia's river basins. SFEWS is a full-stack system that ingests live telemetry from IoT edge sensors deployed across Ethiopian river basins, computes flood risk in real time, and provides an intelligent AI Command Center for emergency operators. The platform combines hardware simulation, machine learning, and large language models to deliver data-backed situational awareness and evacuation recommendations. --- ## ✨ Key Features - **🛰️ Real-Time Telemetry** — Live water level and rainfall data streamed via WebSockets to all connected dashboards instantly. - **🤖 AI Command Center** — Natural language chat interface powered by GPT-4o. Ask questions like *"What is the flood risk at Awash River?"* and get a data-backed emergency report synthesized from live sensors and weather forecasts. - **🔄 Fail-Safe LLM Routing** — Automatic fallback across three AI providers (Azure OpenAI → DeepSeek → Groq) to guarantee zero downtime during presentations or production use. - **🌦️ Weather Integration** — Upstream rainfall forecasts pulled from OpenWeather API and correlated with sensor readings. - **📍 Smart Geocoding** — Converts natural language locations (e.g., *"Abay Bahirdar"*) into precise coordinates using LocationIQ. - **⚡ Event-Driven Architecture** — No polling. The backend pushes updates to all frontends the instant new data arrives. - **📊 Risk Classification** — Every reading is classified into `Safe`, `Warning`, or `Critical` tiers with computed flood probabilities. --- ## 🏗️ Architecture ```text ┌──────────────┐ ┌──────────────────────┐ ┌──────────────────┐ │ IoT Sensors │────▶│ FastAPI Backend │────▶│ React Frontend │ │ (Simulator) │ │ │ │ (Command Center)│ └──────────────┘ │ ┌────────────────┐ │ └──────────────────┘ │ │ PostgreSQL │ │ ▲ │ │ (Neon Cloud) │ │ …