Mapping the Spatial Distribution of Poverty

This report shows how Indonesia can combine satellite imagery and geospatial data with machine learning to create maps showing areas of poverty, which could help shape targeted support for vulnerable communities.

Based on a collaborative study with Statistics Indonesia (BPS) and World Data Lab, the report shows how Indonesia can develop a grid-like poverty map. Computer vision and machine learning can help identify roads, buildings, and vegetation, and that information can be harnessed to estimate poverty levels. The report details poverty reduction efforts in the diverse archipelago and shows how geospatial data can help improve social protection and resource allocation.