Geospatial micro-estimates of slum populations in 129 Global South countries using machine learning and public data

Abstract. Slums are a visible manifestation of poverty in Global South countries. Reliable estimation of slum populations is crucial for urban planning, humanitarian aid provision, and improving well-being. However, large-scale and spatially explicit mapping is still lacking due to inconsistent methodologies and definitions across countries. Existing datasets often rely on government statistics, lacking spatial continuity or underestimating slum populations due to factors such as city image and privacy concerns. Here, we develop a standardized bottom-up approach to estimate slum populations at the grid-cell level (∼6.72 km resolution at the equator) for 129 Global South countries in 2018. Leveraging the Sustainable Development Goal 11.1 framework and machine learning, our estimation integrates publicly accessible household-based surveys, satellite imagery, and gridded population data. The resolution is thus jointly determined by the model input patch size and the spatial resolution of Landsat imagery. Our models explain 82 %–95 % of the variation in within-country spatial prediction and 45 %–91 % in country-level holdout validation, corresponding to median R2 values of 0.89 and 0.85, respectively. These results indicate strong predictive performance and remain broadly comparable with, and in some cases higher than, previously reported benchmarks. To our knowledge, this is the first comprehensive geospatial inventory of slum populations across Global South countries. Although not designed for site-specific operational applications, the dataset provides valuable insights for national and regional urban sustainability assessments and further research on vulnerable populations. The maps of slum populations and their local shares in Global South countries are available in the Zenodo repository at https://doi.org/10.5281/zenodo.13779002 (Li et al., 2025).