Machine learning and GIS project predicting deforestation risk in Embobut Forest, Elgeyo Marakwet County, Kenya
# Embobut Forest Deforestation Prediction (Elgeyo Marakwet, Kenya)
## Overview
This project uses Machine Learning and GIS data to predict areas at risk of deforestation in the Embobut Forest, located in Elgeyo Marakwet County, Kenya.
## Project Workflow
1. **Data Acquisition (Google Earth Engine)** – Download layers like:
- Hansen Global Forest Change (forest loss year, tree cover)
- Sentinel-2 NDVI composites
- SRTM elevation
- Distance to roads and settlements (OSM)
2. **Preprocessing** – Convert raster data into tabular format.
3. **Model Training** – Train ML model to classify deforestation risk.
4. **Web App** – Flask + Leaflet interface to visualize and interact with predictions.
## Region of Interest (ROI)
Embobut Forest (approx. 35.2°E to 35.7°E, 0.9°N to 1.4°N)
## Setup
```bash
pip install -r requirements.txt
python train_model.py
python app.py
```
## Author
Ruto (Embobut Forest Deforestation Prediction )