How can you trust OpenStreetMap?

The most voiced criticisms about OpenStreetMap are that the data is not complete and that the data is not reliable. ‘Completeness’ and ‘reliability’ are relative measures; my map is more complete than yours for my personal definition of completeness. If your requirements for completeness are all the motorways in the UK, or the footpaths of London, or the rail network of the UK, or the Tube stations in London, then OpenStreetMap’s data will fit your needs.

The reliability argument goes like this; ‘How can I trust OpenStreetMap maps. They are volunteers on bikes with cheap GPS, Navteq have vans with expensive GPS.’ Fair enough?

OpenStreetMap volunteer, Dair Grant does not think so. After completing mapping his local town of Haywards Heath, he set about comparing the quality of OpenStreetMap’s data against that of Google Maps, supplied by TeleAtlas. The findings of his study can found here.

Dair found 89 errors in Google Maps, based on the elements that appear in both OpenStreetMap and Google’s Maps (Google do not include bridleways OSM do, for example). These included incorrect junctions, missing roads, incorrect geometries and roads that simply did not exist. To confirm that the ‘errors’ were not on the part of OpenStreetMap, Dair went back out and resurveyed the areas he had pinpointed as containing errors.

Dair’s work clearly demonstrates what most OpenStreetMap volunteers know already - that for areas where OSM’s data exists, it is usually of a high quality. Why would someone give up their free time to make a map and not do it properly? There are no ‘Friday afternoons’ or ‘lazy Mondays’ for volunteers - if you can’t be bothered, you just go home. Conversely, large companies spend massive amounts of money managing and motivating their staff and then making sure that their work is up to scratch.

Of course, volunteers can make mistakes too, but their motivations for revealing their mistakes are different to employees. There is little point in an OpenStreetMap volunteer covering up a mistake they made. All of OpenStreetMap’s data is open, so if you see a mistake in the map, you can access the raw data and fix it. Furthermore, many volunteers make their edits public, meaning that anyone can see the OSM name of the person who collected the data. You earn bad OSM points if the community catch you out. Compare this to the attitude of an employee of a large company. There is an entire management infrastructure dedicated to making sure the employee does their job properly. But unless that infrastructure is working 100% effectively, mistakes are going to slip through. And what is the point in the individual reporting their misspelling or wrong attribution of a road? Furthermore, what do you, the consumer do when the errors reach your rung on the supply chain? The large mapping companies have realised the potential for crowd sourcing data - but they are applying a thin layer on top of their system. They are allowing bug-reporting - they are not exposing the entirety of their software and hardware systems, mapping methodology and management techniques for the inspection of the world. This is why they are borrowing from the innovations of the open community and why the software and ultimately the data of the open community will be superior to that of proprietary producers.

October 13th, 2007 - Posted by in cloudmade, geodata, openstreetmap | |

Leave a reply

You must be logged in to post a comment.