I'm using OSM as the database rather than create another silo of data because I get fed up of people being gatekeepers and I have little interest in being one. That means people can improve the data themselves although I haven't yet fully automated the pipeline to bring data in from OSM.