OpenStreetMap has better coverage of Cambridge than Google Maps. Google Maps combines mapping data from many sources. Why doesn’t it make use of OpenStreetMap?
Nigel Deakin pointed out to me that OpenStreetMap has better coverage of Cambridge than Google Maps. Here’s part of Cherry Hinton in both maps (click on the thumbnails for larger images):
Notice that Open Street Map has data on land use (lakes, green space), cycle paths, footpaths, and notable buildings (Brookfields Hospital, St Bede’s School, Cherry Hinton Hall). Both maps have minor mistakes: for example, OpenStreetMap is missing half of the semi-circular road leading to Cherry Hinton Hall. On the other hand, Google Maps shows the cycle path down the back of the Next Generation Club as a road, and is missing half of The Paddocks and the connection from Laundry Lane to Derwent Close.
Google Maps combines mapping data from many sources. Why doesn’t it make use of OpenStreetMap?
Maybe there’s a licensing issue? The OpenStreetMap licence for data is the Creative Commons Attribution-ShareAlike 2.0 licence, which shouldn’t cause Google any problems. The licence for software is the GNU GPL, but I can’t imagine that causing any trouble, since surely Google will proceed by translating the data into their own geographical information system?
Maybe there’s concern about the accuracy of the data? I believe the high quality of the Cambridge data is owed to Fulbourn resident and cycle campaigner David Earl. But even Google doesn’t have the resources to check the provenance, trustworthiness and accuracy of each piece of data from OpenStreetMap; it needs some kind of automated system. But I’m sure something could be set up. The simplest approach would just be an “OpenStreetMap” view to go with the current “Map”, “Satellite” and “Terrain”. Or Google Maps could do an automated check of the OpenStreetMap data against the purchased data to get a rough idea of the quality of OpenStreetMap’s road data in a particular area, and use that as an input to the decision as to whether to use the other kinds of data.
Maybe no-one has got round to it yet. In which case, here’s a free idea for a 20% project for any Google staff reading this blog.