[OSM-dev] islands and coastlines misaligned

Jon Burgess jburgess777 at googlemail.com
Thu Aug 14 20:00:35 BST 2008


On Thu, 2008-08-14 at 17:18 +0100, Andy Allan wrote:
> Islands and purple state boundaries, I believe it is. So given that
> both of them are auto-generated, one of the sources must have been
> off.
> 
> Cheers,
> Andy
> 
> On Thu, Aug 14, 2008 at 4:37 PM, Mikel Maron <mikel_maron at yahoo.com> wrote:
> > any reason why the coastlines and not aligned with the island land areas in
> > the Great Lakes?
> >
> > http://www.openstreetmap.org/?lat=46.989&lon=-90.777&zoom=11&layers=B00FTF
> >
> > -Mikel

If you turn on the data layer it shows that the rendered tiles match the
underlying data in the OSM DB so it isn't an issue with the rendering.
The coastlines appear to be consistent with the tiger import and the
Yahoo imagery. It looks like the county data is at fault. 

It seems the county data came from:
http://www-atlas.usgs.gov/mld/countyp.html

It looks like the points in the DB match the shapefile. The data
projection data they specify (NAD83/GRS80) seems to be an exact match
for WGS84. 

The error seems to be a shift of around 500m. The FAQ for the county
data has a section for data accuracy but the answer is blank. 

	Jon






More information about the dev mailing list