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

This dataset provides 100-meter gridded estimates of slum prevalence for four major cities in Kenya (Nairobi, Mombasa, Nakuru, and Kisumu), derived from a Random Forest classifier trained on the building level. Each grid cell contains information on the number and area of predicted slum buildings, the total building stock, and corresponding ratios. The dataset is intended to support urban studies and slum mapping research, offering open-access spatial data that can be compared across cities and integrated with other socioeconomic or environmental indicators. Building footprints are sourced from the Google–Microsoft Open Buildings dataset (harmonized by VIDA), and the urban extents follow the Worldwide Delineation of Multi-Tier City-Regions。