[OSM-talk] Statistics on planet.osm.gz

Joerg Ostertag openstreetmap at ostertag.name
Thu Apr 13 09:19:38 BST 2006

The script with which I do these stats is geoinfo.pl from the gpsdrive 
package. It's a pre check before inserting them into the gpsdrive database.
The output of some is already availabel. I think writing the information also 
as *.html output isn't too hard.


> I would like a cgi-bin which creates a short web-site showing:
> - all nodes that lies *exactly* on top of each other: double inserts
  some of them already output already available
> - all line segments that have the very same nodes as start and end: double
> inserts again
  output already available
> - line segments with equal from and to node
  output already available
> - All sets of ways that share a common name and which are chained together
> (where one way end in the start of the other): misuse of the way feature /
> TIGER import ways
  can be done, a little more complicated, but I tiink I can do it.

> - ways without line segments
  should be easy
> - anything with id=0
  output already available
> ...
> In case of only "warnings" (it could be done by purpose to put two line
> segments on top of each other), special key/value pairs could be used to
> mark the objects as "confirmed".
> Of course, it would be really helpfull to include an <a
> href="openstreetmap.org/...">show in map</a> link after each report ;-).

I'll think about this.

I also was thinking about creating automated delete /merge requests



More information about the talk mailing list