# What kind of roads on a GR

###### tags: `archive`

List of GR's and their relation IDs:
https://wiki.openstreetmap.org/wiki/WikiProject_Belgium/Long_Distance_Walking_Routes#GR_-_Main_routes

## Get all the data from the GR

For example, GR5 in Flanders is [relation 3121667](https://openstreetmap.org/relation/3121667)

So we can query on overpass ([link](http://overpass-turbo.eu/s/11df)):

[out:xml][timeout:25];
(
  relation(id:3121667);
);
out meta;
>;
out meta qt;

## GR with minimal rendering:

We can render the data a little prettier (example: grey fro roads, dashed blue for paths).
[Link](https://overpass-turbo.eu/s/1jID) The query is limited to what you see on the map (as requested by GR VZW)

[out:xml][timeout:25];
(
  relation(id:3121667)->.groute;
);
way(r.groute);
out geom({{bbox}});

{{style:

/* define general layout of road classes */

way{
  color:grey;
}

way[highway=path],way[highway=footway],way[highway=cycleway],way[highway=bridleway]{
  dashes:5,5;
  color:blue;
}
way[highway=track]{
  dashes:10,10;
  color:blue;
}
}}

## Summary stats for the entire GR

Thanks to Midgard at https://help.openstreetmap.org/questions/77931/overpass-stats-per-value (see also) https://dev.overpass-api.de/blog/loop_and_group.html

We can generate stats for every combination of highway and surface type ([link](http://overpass-turbo.eu/s/11ga)) ([link bbox version](https://overpass-turbo.eu/s/1jJu)):

[out:csv(highway,surface,number,length)][timeout:25];
relation(id:3121667)->.groute;
way(r.groute);

for->.hw (t["highway"]) {
  for.hw (t["surface"]) {
    make stat highway=hw.val, surface=_.val, number=count(ways), length=sum(length());
    out;
  }
}

This info can then just be copy-pasted to Excel (mind . for , in some versions!)

Making some assumptions, and grouping some data this gives us this little graph: