Résumé : Spatial data on Low-and-Middle-Income-Country (LMIC) cities, and deprived areas within cities, are often not readily available in support of local and global information needs. To address this information gap, we propose the systematic semi-automated SLUMAP framework that provides policy-relevant information on deprived urban areas in Sub-Saharan Africa (SSA), based on free open-source software (FOSS). First, we assess user needs for spatial information on deprivation (ranging from local communities to global research and policy support). Second, we show how free or low-cost image datasets can be used for mapping the location of deprived areas at the city scale and providing an overall assessment of their spatial patterns. This is implemented as a grid-based approach using machine learning and assessing the contribution of a large number of spectral and spatial features derived from open or low-cost imagery. Third, we show how higher (spatial and spectral) resolution data can provide a detailed characterization of such areas, with a GEOBIA/machine-learning workflow and deep learning techniques. We illustrate the experiments and results on the city of Nairobi (Kenya)and discuss transferability to SSA cities.