Data Transparency

Our predictive modeling pipeline is engineered exclusively on public geospatial and satellite datasets, ensuring that our methodology can be audited, reproduced, and scaled by the broader scientific community.

Hansen Global Forest Change

University of Maryland / GEE30m (resampled to 1km)

Baseline forest cover (treecover2000) and historical forest loss labels.

Open License

GFW Integrated Alerts

Global Forest WatchDaily/Weekly signals

Near-real-time deforestation evidence utilized as predictive features.

Open License

Sentinel-2 Multispectral

European Space Agency (ESA)10m-60m

Spectral bands (B2-B12) and derived indices including NDVI and NBR.

Open License

SRTM Digital Elevation

NASA / USGS30m

Terrain slope and accessibility pressure features derived from elevation.

Open License

OpenStreetMap Roads

OSM ContributorsVector

Road networks to compute Euclidean distance to nearest infrastructure.

Open License

Open Research Repository

The processed 1 km resolution analytical master tables for the K’Bang and Mang Yang pilot areas, which merge all aforementioned sources, are available directly within our GitHub repository.

Available for non-commercial academic research and replication.