Building-Level Slum Prevalence Indicators Aggregated to 100-Meter Grids in Kenyan Cities

This dataset provides 100-meter resolution maps of slum prevalence across four major Kenyan cities: Nairobi, Mombasa, Nakuru, and Kisumu. The estimates are derived from a Random Forest (RF) model trained on building-level features, including morphological, spectral, distance-based(proximity), and topographic features obtained from open data. Each grid cell reports the number and area of predicted slum buildings, together with total building counts and corresponding ratios. The dataset and RF model offer an open, scalable, transferable approach on detecting slum areas and futhur contribute to research on urban poverty and sustainability. *For more details, please refer to the publication: A transferable and interpretable approach to slum mapping using building morphometrics and optical imagery (https://doi.org/10.1080/15481603.2026.2649308)