Asfemi/NFAS

Nigeria Flood Alert System, enriching grassroot communities and farmers with needed timely flood information to prevent flooding losses # NFAS Nigeria Flood Alert System, enriching grassroot communities and farmers with needed timely flood information to prevent flooding losses # Flood Sentinel Nigeria A multilingual flood alert and advisory system that transforms complex flood risk data into clear, actionable warnings for communities across Nigeria. ## Problem Nigeria already has flood forecasting data (e.g., NIHSA, Google Flood Models), but: - It does not reach local communities effectively - It is not easy to understand - It is not actionable - It is not localized or multilingual This leads to preventable loss of lives and property. ## Solution Flood Sentinel Nigeria bridges this gap by: - Accepting a user's **Local Government Area (LGA)** - Retrieving flood risk data - Converting it into **plain-language advice** - Returning **English** plus **one regional language** per LGA (South West → Yoruba; South East / South South → Igbo; North / North Central → Hausa) - Formatting it into **SMS-ready alerts (≤160 characters)** ## MVP Scope ### Features - LGA-based flood risk query - AI-generated plain-language alerts - Bilingual SMS output: English + auto-selected regional language for the LGA’s state - SMS-ready formatting - Lightweight web interface ### Not Included (Phase 2) - Live satellite data integration - Direct NIHSA API connection - Mobile app - Real-time push notifications --- ## How It Works 1. User selects or inputs their LGA 2. System retrieves flood risk level from dataset 3. AI generates: - Risk explanation - Safety advice 4. Output is: - Translated into 4 languages - Condensed into SMS format --- ## Tech Stack - **Frontend:** Web (React / Next.js or similar) - **Backend:** Node.js / Python (API layer) - **AI Engine:** Google Gemini API - **Data Source:** Curated flood-risk dataset (LGA-based) - **Deployment:** Vercel / Cloudflare / Firebase --- ## Project Structure /frontend /backend /data /docs ## 📊 Data Model (MVP) ```json { "lga": "Lokoja", "state": "Kogi", "ri …