We analyzed thousands of satellite images to automatically identify buildings missing from OpenStreetMap. Using European Space Agency’s Sentinel-2 imagery from Astro Digital, we analyzed the images to first identify objects that look like buildings then we overlay these suspects on OSM buildings layer to find candidates for buildings that are not labeled in OSM. With an up to date, global imagery source, this approach could help keep OSM updated by adding new buildings as they appear and flagging buildings that are no longer there.