Forest Intelligence Platform v0.1 Beta — All-Weather Forest Analytics

Forest Intelligence Platform is a map-driven analytics system that estimates forest metrics from satellite features and exposes them through a practical dashboard + API workflow.

It is designed to demonstrate geospatial system design, satellite feature engineering, and production-style backend/frontend integration—while remaining usable for real polygon analysis.

---

📋 Table of Contents

• 🚀 What This Project Does
• 🧠 How It Works (Simple Explanation)
• 📦 Download & Run
• 🧭 Main Commands
• 🧪 Typical First-Time Usage
• ⚙️ Requirements
• 🛡️ Reliability & Data Notes (Plain English)
• ✨ Features
• 🗂️ Repository Structure
• 👤 Author

---

🚀 What This Project Does

This project allows you to:

✅ Draw a polygon and receive forest analytics for that area

✅ Estimate tree density and tree count based on satellite-derived features

✅ Compute forest health and risk indicators

✅ View forecast trends and secondary analytics in one dashboard

✅ Run locally with FastAPI + React + PostGIS

---

🧠 How It Works (Simple Explanation)

At a high level:

🛰️ Sentinel features are prepared through ingestion + processing + extraction

🗺️ You draw a polygon on the map

📊 Backend computes area-scoped metrics from DB/features and ML utilities

🔄 Frontend polls pipeline status first, then requests metrics when ready

You do not need GIS expertise to use it—the dashboard coordinates this workflow for you.

---

📦 Download & Run

Step 1 — Clone and enter project

git clone https://github.com/HetMistri/forest.git
cd forest

Step 2 — Start backend

cd backend
uv sync
cp .env.example .env
uv run uvicorn api.main:app --reload --host 0.0.0.0 --port 8000

Step 3 — Start frontend (new terminal)

cd frontend
npm install
npm run dev

App URLs:

• Frontend: http://127.0.0.1:5173
• Backend Docs: http://127.0.0.1:8000/docs

---

🧭 Main Commands

Backend

• uv run uvicorn api.main:app --reload --host 0.0.0.0 --port 8000
• uv run python main.py
• uv run pytest

Frontend

• npm run dev
• npm run build
• npm run preview

Pipeline

• python scripts/run_backend_pipeline.py
• cd backend && uv run python scripts/run_smoke_pipeline.py

Docker (Backend + PostGIS)

• docker compose up --build
• docker compose down -v

---

🧪 Typical First-Time Usage

1. Start backend and frontend
2. Open dashboard in browser
3. Draw polygon on map
4. Wait for pipeline status to become ready
5. Review forest metrics + forecast + risk outputs
6. Run smoke pipeline if polygon has no prepared features

---

⚙️ Requirements

To run locally:

🐍 Python 3.11+

📦 Node.js + npm

🗄️ PostgreSQL/PostGIS (or Docker compose service)

🛰️ Earth Engine credentials for full ingestion pipeline

---

🛡️ Reliability & Data Notes (Plain English)

• Metrics are model/data-derived estimates, not manual field census numbers.
• In strict mode, backend may return 503 when data is not yet available.
• DEMO_CACHE_ENABLED can return cached responses for known polygons.
• Best results require prepared feature data in forest_features for the selected area.

---

✨ Features

🌧️ All-weather oriented pipeline support (including SAR-driven features)

📉 Health and risk scoring from NDVI/NDMI/SAR summaries

📈 Forecast and secondary analytics endpoints

🧩 API + frontend integration with request logging and retry logic

🐳 Docker-ready local backend + PostGIS stack

---

🗂️ Repository Structure

backend/                      FastAPI API + services + SQL + pipeline modules
frontend/                     React dashboard
backend/services/ml/          ML utilities (density, health/risk, forecast)
data/                         Raw/processed artifacts and metadata
scripts/                      Root pipeline runner
docker-compose.yml            Local PostGIS + backend services
README.md                     Main project guide

---

👤 Author

Het Mistri

LinkedIn: https://www.linkedin.com/in/het-mistri-7a52a533a/

GitHub: https://github.com/HetMistri

---

For module-level setup details, see:

• backend/README.md
• frontend/README.md
• backend/services/ml/ml_README.md