Mapbox Satellite is used by designers, developers, cartographers, and mappers, all of whom depend on clear and up-to-date imagery. In order to assess the quality of this massive set of data — at higher zoom levels there are over 1.7 quadrillion pixels in Mapbox Satellite — we needed an efficient yet robust solution.

By using random sampling, interpolation, and a suite of evolving quality detection algorithms, we’ve developed a Satellite Health Index (SHIdx) that highlights areas of poor imagery quality. This workshop will walk through how we use Open Source tools and Mapbox APIs to implement SHIDx at scale in order to strategize imagery refreshes. I’ll also touch on how we are using feedback from the OpenStreetMap and HOT communities, as well as our user base at large to improve our quality assessment and imagery update efforts.