Projections suggest that most of the global growth in population in the next few decades will be in urban centres in Asia and Africa. Most of these additional urban residents will be concentrated in slums. However, government documentation of slums is incomplete and unreliable, and many slums remain undocumented. It is necessary to employ creative methods to locate and sample these understudied populations. We used satellite image analysis and fieldwork to build a sample of Indian slums. We show that living conditions vary along a wide-ranging continuum of wellbeing; different points correspond to different policy needs. We also show that most variation in conditions is due to differences across rather than within neighbourhoods. These findings have important implications for urban policy. First, satellite data can be a useful tool to locate undocumented settlements. Second, policy must be appropriately nuanced to respond to wide-ranging needs. Finally, variation patterns suggest that policies should be targeted at the neighbourhood rather than the individual level.
