[OSM-talk] Google Maps comparison

[Corey Burger]

On 10/10/07, Dair Grant <dair at refnum.com> wrote:
> Hi,
>
> I recently finished mapping the town I live in, and thought it
> would be interesting to do a detailed comparison with Google
> Maps (or really, the TeleAtlas data it is based on).
>
> Once the OSM version was finished, I reviewed each street to
> check it against the equivalent in GM. You can see the results at:
>
>      <http://www.refnum.com/osm/gmaps.html>
>
> Although the missing road figures are inflated by a "new" (2004)
> housing estate, the overall error rate is surprisingly high.
>
>
> This should also be confirmation (if anyone was in any doubt)
> that entering data into OSM by tracing a commercial source is a
> really really bad idea.
>
> Many of these errors will probably go uncorrected for years, and
> it would be extremely easy to insert deliberate errors without
> affecting the usability of the map.
>
> It would also be trivial to re-survey an area to detect these
> kind of errors, then use that list of corrections to check for
> identical mistakes in a suspect map.
>
>
> -dair

Dair,

That is great stuff. It is the first qualitive analysis of "complete"
data from us and GMaps. We should be promoting this more.

Corey