Transforming pastoral resilience through satellite intelligence, machine learning, and logistics optimization
Live Demo • Documentation • Screenshots • Team • Contributing
| Section | Description |
|---|---|
| About | What is መስክAI and why it matters |
| Features | Complete feature breakdown with screenshots |
| Tech Stack | Technologies powering the platform |
| Architecture | How components work together |
| Team | Meet the contributors |
| Quick Start | Run locally in 2 minutes |
| Security | Security practices |
| Roadmap | Future plans |
መስክAI is an enterprise-grade, AI-powered platform designed specifically for Ethiopian pastoral communities facing climate-induced forage scarcity. By leveraging Sentinel-2 satellite imagery, Google Earth Engine, machine learning models, and optimization algorithms, መስክAI delivers real-time forage predictions, livestock management insights, and intelligent feed dispatch routing.
Ethiopia's pastoral regions support millions of livestock and livelihoods, yet face:
- Unpredictable drought cycles causing massive livestock mortality
- Lack of real-time forage data for migration decisions
- Inefficient emergency feed logistics reaching affected areas too late
- Fragmented information across weather, satellite, and ground data
መስክAI provides a unified command center with:
| Capability | Impact |
|---|---|
| Satellite NDVI Monitoring | Real-time vegetation health across all zones |
| AI-Powered Predictions | 60-day forage outlook using Gemini AI |
| Climate Analytics | Temperature, precipitation, drought indexing |
| Route Optimization | Clarke-Wright CVRP algorithm for feed dispatch |
| Executive Briefs | Auto-generated decision reports for policymakers |
Full-screen cinematic hero with parallax effects, gradient text logo ("መስክ"), and compelling call-to-action buttons. Shows live monitoring status and coverage statistics.
Interactive feature cards showcasing the three core capabilities: Sense (Sentinel-2 NDVI + Open-Meteo rainfall), Forecast (Ensemble models for forage risk at 15/30/45/60 days), and Dispatch (Clarke-Wright routing for feed logistics).
Complete command center overview featuring:
- KPI Header -- Avoidable loss ($7.6M), zones monitored (10), critical alerts (2)
- Statistics Grid -- 6 animated cards: Zones, Critical, Feed Deficit, At Risk, Mean NDVI, Loss at Risk
- Projection Timeline -- Interactive 60-day slider (Today / +15d / +30d / +45d / +60d / Seasonal)
- Situational Awareness -- Live Leaflet map with district overlays and priority zone panel
Meteorological intelligence dashboard powered by Open-Meteo global coverage:
- Correlation Chart -- 7-Day Cumulative Rainfall, Max Temperature, Drought Severity Index across all zones
- Zone Weather Cards -- Per-district conditions (Borena Zone, Afar Zone, Siti/Shinile) with live status
- Metrics -- Rainfall totals, temperature readings, weather icons per zone
Climate Data Sources: Open-Meteo API with real-time rainfall, temperature, humidity, and drought severity indices.
Sentinel-2 vegetation intelligence via Google Earth Engine:
- Weather-Assimilated Model -- Rangeland phenology calibrated with Sentinel-2
- Sensor Metadata -- COPERNICUS/S2_SR_HARMONIZED, composite dates, raw DN values
- NDVI Forecast Chart -- Historical observed NDVI + Ensemble forecast + Moving Average + Holt's Exponential Smoothing + Polynomial Regression
- Zone Selector -- Dropdown to switch between monitoring zones
Satellite Features: 10m resolution multispectral imagery, NDVI calculation, cloud-free compositing, trend analysis.
Capacity-aware feed dispatch using Clarke-Wright Savings CVRP algorithm:
- Feed Depots -- 5 strategic locations (Yabelo, Adama, Dire Dawa, Hawassa, Jijiga) with stock levels and fleet availability
- Network Summary -- 9 active routes, 3,112 km total distance, 1,181 liters estimated fuel
- Dispatch Schedule Table -- Depot, target zone, feed load, vehicle type, stops, distance, duration, fuel, status
- Validation Badge -- "2/9 Feasible Routes validated"
Optimization Constraints: Truck capacity (6T-20T, 10T-10T fleets), depot stock checks, multi-stop routing.
Executive decision support briefs powered by Google Gemini AI:
- Zone Brief Header -- Borena Zone, +30 day horizon, date, Rules Engine source
- Priority Warning Badge -- Color-coded alert (Warning/Critical/Normal)
- Executive Summary -- Plain-language prediction ("moderate pasture stress... decisive action by day 15 could save ~1,576 head")
- Livestock Saved Comparison -- Action-by +15d / +30d / +45d / +60d with "BEST BY DAY 15" highlight
- Print Brief Button -- Export-ready formatting
AI Fallback: Deterministic rules engine when Gemini API is unavailable.
Herd pressure and emergency feed need assessment:
- Summary Cards -- Feed Needed (4,066 metric tons), Livestock at Risk (1,299,591 head), Asset Loss at Risk ($7.59M USD)
- Action Date Table -- Per-district livestock saved by intervention timing (+15d / +30d / +45d / +60d) with best window recommendation
- Zone Breakdown -- Borena Zone, Afar Zone 1, Siti (Shinile), Bale Lowlands with TLU-based deficit estimates
- Impact Metric -- "+30d -> 25,532 head saved" banner
Data Source: Literature-aligned zone estimates (not live census), modeled from site NDVI risk and herd pressure.
| Technology | Version | Purpose |
|---|---|---|
 |
19.x | UI Library |
 |
5.7.x | Type Safety |
 |
6.x | Build Tool |
 |
4.x | Styling |
 |
11.x | Animations |
 |
1.9.x | Maps |
 |
2.x | Charts |
| Technology | Version | Purpose |
|---|---|---|
 |
22.x | Runtime |
 |
5.x | Server Framework |
 |
API | Satellite Data |
 |
API | AI Analysis |
 |
API | Weather Data |
+-- Color Palette
| +-- Field Green (#22C55E) -- Healthy vegetation
| +-- Signal Orange (#F97316) -- Warning/Attention
| +-- Critical Red (#EF4444) -- Danger/Urgent
| +-- Sky Blue (#0EA5E9) -- Information/Water
| +-- OK Green (#84CC16) -- Success/Normal
|
+-- Glass Morphism
| +-- Backdrop Blur (12px)
| +-- Semi-transparent backgrounds (rgba)
| +-- Subtle borders (1px solid rgba)
|
+-- Animations
+-- Fade Up (entrance)
+-- Scale In (modal)
+-- Gradient Shift (hero)
+-- Shimmer (loading)
+-- Pulse Ring (live indicators)
+-------------------------------------------------------------------------+
| CLIENT LAYER |
| +----------+ +-----------+ +----------+ +----------------------+ |
| | React | | Tailwind | | Framer | | Recharts | |
| | 19 + TS | | CSS 4 | | Motion | | Visualization | |
| +----+-----+ +-----+-----+ +----+-----+ +-----------+----------+ |
| +--------------+----+-------------------+ |
| V V |
| +-------------------------------+ |
| | Leaflet Maps | |
| | (6 Basemap Layers) | |
| +---------------+---------------+ |
+------------------------------+----------------------------------------+
| HTTP/REST
+------------------------------V----------------------------------------+
| API LAYER |
| +---------------------------------------------------------------+ |
| | Express.js Server | |
| | +----------+ +----------+ +----------+ +---------------+ | |
| | | /climate | |/satellite| |/logistics| | /ai-brief | | |
| | +----+-----+ +----+-----+ +----+------+ +-------+-------+ | |
| +-------+------------+------------+---------------+--------------+ |
+----------+------------+------------+---------------+--------------------+
| | | |
+----------V------------V------------V---------------V--------------------+
| SERVICE LAYER |
| +------------------+ +------------------+ +---------------------+ |
| | geeService | | weatherService | | aiAnalyzer | |
| | (Sentinel-2) | | (Open-Meteo) | | (Gemini AI) | |
| +------------------+ +------------------+ +---------------------+ |
| +------------------+ +------------------+ +---------------------+ |
| | forecasting | | feedEstimator | | routeOptimizer | |
| | (NDVI Predict) | | (Requirements) | | (CVRP Algorithm) | |
| +------------------+ +------------------+ +---------------------+ |
+-------------------------------------------------------------------------+
| | |
+----------V------------V------------V------------------------------------+
| EXTERNAL APIS |
| +------------+ +------------+ +------------+ |
| | Google EEE | | Open-Meteo | | Gemini AI | |
| | (Satellite)| | (Weather) | | (LLM) | |
| +------------+ +------------+ +------------+ |
+-------------------------------------------------------------------------+
| Role | Name | Location | Focus Area |
|---|---|---|---|
| Frontend Lead | Zeamanuel Million | Addis Ababa, ET | React, TypeScript, UI Components |
| UI/UX Designer | Bekan Seifu | Addis Ababa, ET | Design System, Glass Morphism, Animations |
| AI/ML Engineer | Elshaday Habtamu | Addis Ababa, ET | GEE Integration, NDVI Models, Gemini AI |
| Backend Engineer | Abdi Megersa | Addis Ababa, ET | Express.js, APIs, Route Optimization |
| Product/Other | Dawit Getachew Tariku | Addis Ababa, ET | Documentation, Testing, Coordination |
- Node.js >= 18.0.0
- npm >= 9.0.0
- Git
# Clone the repository
git clone https://github.com/Hope0351/mesek-AI.git
cd mesek-AI
# Install dependencies
npm install
# Set up environment variables
cp .env.example .env
# Edit .env with your API keys
# Start development server
npm run dev| Variable | Description | Required |
|---|---|---|
GEMINI_API_KEY |
Google Gemini AI API key | Yes |
GEE_PROJECT_ID |
Google Earth Engine project ID | Yes |
GEE_PRIVATE_KEY |
GEE service account key | Yes |
OPEN_METEO_URL |
Open-Meteo base URL | No (has default) |
PORT |
Server port | No (default: 3000) |
npm run dev # Start development server (port 3000)
npm run build # Build for production
npm run start # Start production server
npm run lint # Run ESLint
npm run preview # Preview production builddocker build -t pastureai .
docker run -p 3000:3000 --env-file .env pastureai
| Category | Count |
|---|---|
| React Components | 15+ |
| Pages/Routes | 7 |
| Backend Services | 8 |
| CSS Design System | ~940 lines |
| API Endpoints | 12+ |
| External Integrations | 4 |
| Feature Screenshots | 8 (unique pages) |
- No hardcoded secrets -- All credentials via environment variables
.envin.gitignore-- Prevents accidental secret commits- API Key validation -- Server-side verification before external calls
- Input sanitization -- All user inputs validated and sanitized
- CORS configuration -- Restricted cross-origin access
- Security headers -- Helmet.js middleware ready
- Multi-language support — Amharic (አማርኛ) UI with EN/አማ toggle; Oromo & Somali next
- Multi-language support (Oromo, Somali)
- Offline-first capabilities with service workers
- Push notifications for alert thresholds
- Export to PDF/Excel for reports
- Mobile apps (React Native / PWA)
- SMS/USSD integration for feature phones
- IoT sensor integration for ground truthing
- Community crowdsourcing features
- Federated learning across regions
- Drone imagery integration
- Blockchain for aid transparency
- Pan-African expansion
We welcome contributions! Please see our CONTRIBUTORS.md for details.
- Fork the repository
- Create your feature branch (
git checkout -b feature/amazing-feature) - Commit your changes (
git commit -m 'Add amazing feature') - Push to the branch (
git push origin feature/amazing-feature) - Open a Pull Request
This project is licensed under the MIT License -- see the LICENSE file for details.
- Google -- Earth Engine API & Gemini AI
- Open-Meteo -- Free weather API
- Ethiopian Ministry of Agriculture -- Domain expertise
- Pastoral Communities -- Inspiration and feedback
| Platform | Link |
|---|---|
| GitHub | Hope0351/mesek-AI |
| Contact team via contributor emails above | |
| Location | Addis Ababa, Ethiopia |
Built with care for Ethiopian pastoral communities