[OSM-talk] Statistics on road network length?

Hermann Peifer peifer at gmx.eu
Thu Nov 17 09:26:09 GMT 2011


On 16/11/2011 17:27, Matthias Meißer wrote:
> Thanks Frederik! I think this will do it :)
>
> bye
> Matthias
>

Just to let you know: I am occasionally running my own script which is 
kind of similar to Frederik's Perl script. As OSM input data, I am using 
the .pbf files from [1]

For the Haversine Formula, I followed the explanations at [2]. This 
formula calculates distances between 2 points on a spherical Earth. 
There can be 1% of differences compared to distances measured on an 
ellipsoid. I am using 6367516.477 m as Earth radius, which is the mean 
of a^2/b and b^2/a, where a and b are the semi-major/minor axis of the 
WGS84 ellipsoid.

(As a side remark: the distance between 2 nodes of an OSM way is often 
below 100m, so any considerations about distances on spheres vs. 
ellipsoids are rather of an academical nature.)

Concerning the issue with dual carriageways, I am using a brute-force 
approach and divide the calculated distance by 2 if these conditions are 
true:

highway ~ /^(motorway|trunk|primary|secondary|tertiary)/ &&
  oneway ~ /yes|true|1/

(If someone has a better algorithm: I would be interested.)

Hermann


[1] http://download.geofabrik.de/osm/europe/
[2] http://mathforum.org/library/drmath/view/51879.html



More information about the talk mailing list