FarmPulse

Farm intelligence dashboard for East African smallholder farmers.

FarmPulse combines real-time weather forecasting with AI-powered tree canopy analysis to deliver actionable, context-aware insights — not just data. Instead of showing a farmer "rain is coming" and "12 trees need care" separately, FarmPulse cross-references both to surface recommendations like "rain expected Thursday — ideal conditions to replant the 4 flagged trees."

Built as a technical assessment integrating the WeatherAI API.

Live demo: farmpulse-1.onrender.com
Backend: farmpulse-l2c0.onrender.com

---

Features

• Current conditions — temperature, humidity, wind, UV index, with Gemini AI summary
• 7-day forecast — daily breakdown with precipitation probability
• Auto-detect location — uses WeatherAI's IP geo-detection to find the user's farm region automatically
• Farm tree analysis — upload a drone or satellite image to get tree count, canopy coverage, health breakdown (healthy / needs care / needs replacement), and agronomic recommendations powered by OpenCV + Gemini
• Insight Banner — cross-references weather forecast and tree health to generate combined, actionable farm recommendations
• Annotated overlay — visual output showing detected tree crowns on the uploaded image

---

Tech Stack

Layer	Technology
Frontend	React, Vite
Backend	Node.js, Express
Image handling	Multer (memory storage)
HTTP client	Axios
API	WeatherAI (/v1/weather, /v1/weather-geo, /v1/trees/analyze)
Frontend deploy	Netlify
Backend deploy	Render

---

Architecture

Client (React + Vite)
    │
    │  GET /api/weather?lat=&lon=
    │  GET /api/weather/current
    │  POST /api/trees/analyze  (multipart/form-data)
    ▼
Express Proxy (Node.js)          ← API key lives here only
    │
    │  Authorization: Bearer wai_...
    ▼
WeatherAI API (api.weather-ai.co)

The backend is a deliberate thin proxy — its only responsibility is keeping the API key server-side and forwarding requests to WeatherAI. All UI logic and data composition lives in the frontend.

---

Local Setup

Prerequisites

• Node.js 18+
• A WeatherAI API key — sign up at weather-ai.co

1. Clone the repo

git clone https://github.com/kiki-glow/FarmPulse.git
cd farmpulse

2. Backend

cd server
npm install

Create server/.env:

WEATHER_AI_API_KEY=wai_your_key_here
PORT=3001

Start the server:

npm run dev

Backend runs at http://localhost:3001. Verify with:

curl http://localhost:3001/health

3. Frontend

cd client
npm install

Create client/.env:

VITE_API_URL=http://localhost:3001

Start the dev server:

npm run dev

Frontend runs at http://localhost:5173.

---

API Endpoints (Backend)

Method	Endpoint	Description
GET	/health	Server health check
GET	/api/weather?lat=&lon=&days=	7-day forecast + current conditions
GET	/api/weather/current	Auto-detect location via IP + current weather
POST	/api/trees/analyze	Upload farm image for tree analysis

Tree analysis request

curl -X POST http://localhost:3001/api/trees/analyze \
  -F "image=@/path/to/farm.jpg" \
  -F "location=-1.2921, 36.8219"

Tree analysis response (shape)

{
  "total_tree_count": 84,
  "canopy_coverage_pct": 41.2,
  "confidence_score": 0.87,
  "tree_health": {
    "healthy": 68,
    "needs_care": 12,
    "needs_replacement": 4
  },
  "recommendations": [
    "Consider thinning northern section to improve light penetration",
    "Improve drainage around water source trees"
  ],
  "overlay_image_url": "https://storage.googleapis.com/..."
}

---

Project Structure

farmpulse/
├── server/
│   ├── routes/
│   │   ├── weather.js        # Weather + geo-detection proxy
│   │   └── trees.js          # Image upload + tree analysis proxy
│   ├── middleware/
│   │   └── errorHandler.js   # Centralised error handling
│   ├── .env.example
│   └── index.js
│
└── client/
    ├── src/
    │   ├── api/
    │   │   └── farmpulse.js  # All fetch calls to Express backend
    │   ├── components/
    │   │   ├── Layout.jsx
    │   │   ├── Header.jsx
    │   │   ├── LocationSearch.jsx
    │   │   ├── WeatherPanel.jsx
    │   │   ├── ForecastStrip.jsx
    │   │   ├── TreeAnalysisPanel.jsx
    │   │   └── InsightBanner.jsx
    │   ├── utils/
    │   │   └── weatherIcon.js
    │   └── App.jsx
    ├── .env.example
    └── vite.config.js

---

WeatherAI APIs Used

GET /v1/weather — current conditions and 7-day forecast by coordinates, with Gemini AI summary.

GET /v1/weather-geo — auto-detects caller location from IP, returns weather + geo metadata. Used for the one-click location detection feature.

POST /v1/trees/analyze — accepts a farm image as multipart/form-data. Returns tree count, density per acre, canopy coverage, health breakdown, annotated overlay image, and Gemini-powered agronomic observations.

---

Environment Variables

Server (server/.env)

Variable	Description
WEATHER_AI_API_KEY	WeatherAI API key (prefixed wai_)
PORT	Port to run Express on (default: 3001)

Client (client/.env)

Variable	Description
VITE_API_URL	Base URL of the Express backend

---

Deployment

Backend (Render)

1. Push server/ to GitHub
2. Create a new Web Service on Render, point to the repo
3. Set build command: npm install
4. Set start command: node index.js
5. Add environment variable: WEATHER_AI_API_KEY

Frontend (Render)

1. Push client/ to GitHub
2. Create a new site on Netlify, point to the repo
3. Set build command: npm run build
4. Set publish directory: dist
5. Add environment variable: VITE_API_URL pointing to your Render backend URL

---

Author

Kinya — github.com/kiki-glow