Dataset: A third of sub-Saharan Africa's smallholders cultivate less than half a hectare of land

Farm sizes are highly heterogeneous across sub-Saharan Africa, yet remain difficult to characterize consistently at continental scale. This dataset harmonizes household survey data from 16 countries and uses machine learning to predict the empirical distribution of smallholder farm sizes within 10 km resolution grid cells across the continent. The dataset provides raster layers (.tif) representing the spatial distribution of mean farm size, selected distribution centiles, and the estimated number of farms per grid cell. In addition, we provide a reproducible workflow that can be applied and refined as new or improved data become available.