11 October 2007

Compare OpenStreetMap vs Google Maps Data Sources

An Update:Topologically Integrated Geographic Encoding and Referencing TIGER Data going into OpenSteetMap -
  • with experience with TIGER data, towns and cities have fairly good data, though in more rural locations the data is of less (made up in some places) quality.
  • 'paper roads' can exist - where the roads have been planned but don't physically exist.
  • Poor positional accuracy has however made it extremely difficult to integrate TIGER with advanced technologies and data sources such as GPS, high resolution imagery.
  • The positional accuracy of TIGER road centerline can be improved from an original 100 plus meters using a method called 'conflation'(See Cartography)
Stats on the loading http://dev.openstreetmap.org/~daveh/tiger/stats.html

current stats
"We are processing 6.212597 OSM objects per second at this rate of upload, the entire US will be done in: 27401872 seconds or 317 days or 0.868313 years on Sun Aug 24 12:27:57 2008"

OpenStreetMap vs Google Maps Data Sources - who do you trust?

This is a very interesting topic and compares commercial products (TeleAtlas) with community captured data (OpenStreetMap)

Note: All Maps are to an extent wrong/or out of date, but having a good baseline of cleaned digital data is a very good start.

RefNum has produced set of 'editorial' markups showing the differences and even for a small town there are a number of striking ones.
Haywards Heath Errors - KML Overlay
In detail on Hayward's Heath is a KML Overlay highlighting errors.

Comparing TeleAtlas Google Maps with Open Street Map
has side by side comparison of Google Maps and Open Street Map (MapNik Styled)

"This is an attempt at generating nice map tiles for slippy maps at low zoom levels which can then be combined with OpenStreetMap tiles from higher zoom levels.
The information in the OpenStreetMap database is intended for higher
zoom level and doesn't work well when viewing the whole world or a whole continent."
Jochen Topf

So the future looks good for community based mapping data (open source free mapping).
Commercial digital data capture companies have the edge for coverage at the moment. But the more the people and small companies that join in the faster the benefit to all.
It will not put out business to business mapping but for 'Joe Public'
without the threat of breaking licencing infringements to produce maps for personal/club/internet usage.

Interested in this Project?
Then head to OpenStreetMap



Labels: , , , , , ,


At Thursday, October 11, 2007 4:44:00 pm, Blogger Christopher Schmidt said...

OpenStreetMap data is not 'copyright free' -- it is 'openly licensed' (CC-By-SA).

This is an important distinction, especially given the -SA clause in the current license -- it has affects on what you are legally able to do with the data in some circumstances, in order to ensure that organizations do not 'steal' the data without contributing back to the project.

At Saturday, October 13, 2007 3:03:00 am, Anonymous Anonymous said...


If you have TIGER data which has some kind of semi-systematic problems (test first with some control points) then you can use an affine correction to fix it. Here is one such example, worked out with R and PostGIS.




At Saturday, October 13, 2007 12:43:00 pm, Blogger Simon said...

Can I use Open Street data at my job - ie an environmental consultancy?
Do I need to have a disclaimer/copyright details?


Post a Comment

<< Home