Just let’s put all that Microsoft building data in! The amount of buildings in this and other open data sources is tempting. It’s well possible to top existing 485 Million OSM buildings up to Billion in total.
But did you have a look at what you are going to import here?
Have you checked geometries for validity? What does validity even mean? Are there too low, too small, or too large buildings?
Then the fun begins with solving overlapping objects. Which candidate is more correct?
At least empty spots are trivial to fill. Unless you’re going to replace a recently demolished building in OSM.
At some point you may find yourself asking: “what about quality in OpenStreetMap itself”?
We’ll discuss id’s, creative attributes and the question of all questions: what is a building at all?
No recording is available from this session.
Next up in State of the Map USPrevious talk
TomTom launched MapMetrics, a system to analyze map quality for OSM globally, in August 2021 (https://www.tomtom.com/blog/maps/tomtom-openstreetmaps-mapmetrics-map-editing/). The system uses GPS traces to see where people go, and route against those traces to...