About · Bird Population Decline AI

The idea

Bird populations are sensitive indicators of ecosystem health, and their decline is often tied to habitat loss — deforestation, wetland drainage, and urban sprawl. Those changes are visible from space. Bird Population Decline AI makes that connection tangible: click any point on Earth and get an immediate, index-based readout of vegetation health, water, built-up density, land-cover, and a habitat-health / decline-risk score.

Data sources

Sentinel-2 MSI (ESA / Copernicus) — free multispectral surface reflectance, accessed through the Sentinel Hub Process API on the Copernicus Data Space Ecosystem.
Mapbox GL JS — the interactive satellite basemap.
OpenWeather — live weather conditions for context.

How it's built

The app is intentionally lightweight: a static frontend (plain HTML/CSS/JS with Mapbox and Chart.js) talking to a handful of Python serverless functions that fetch imagery, compute indices with NumPy, render heatmaps with Matplotlib, and build PDF reports. It deploys with zero configuration and no persistent servers. Read the full pipeline on the Methodology page.

Honest limitations

This is a demonstration and educational tool, not a validated scientific instrument. The habitat and risk scores are proxies derived from a single snapshot of vegetation index values — they do not incorporate species data, historical population counts, or ground-truth ecology. Cloud cover, seasonality and imagery availability all affect results. Do not use these numbers for real conservation decisions without expert validation.

Credits

Built with ❤️ using open data and open-source tools. Sentinel-2 imagery © ESA / Copernicus. This project is released under the MIT License.

Satellites for bird habitats

The idea

Data sources

How it's built

Honest limitations

Credits