The Main Roads department of Western Australian recently released quite a few datasets under the Creative Commons Attribution licence. We have explicit permission to use two of these in OSM: the Road Network (MRWA-514) and Road Hierarchy (MRWA-515).

I’m part of a local geogeeks group that meets fortnightly in Perth, and with the help of some people far more knowledgeable than me about spatial stuff I’ve been attempting to derive a comparison between the Main Roads data and what’s already in OSM. That’s a work in progress (eventually it’ll spit out an OSM file containing only correctly-attributed roads that are missing from OSM), but for the time being I’m going down a much more manual route:

1. Download shapefile from the data.wa.gov.au site linked above

2. Use ogr2osm to convert it to OSM format:

python ogr2osm.py RoadNetworkMRWA_514/RoadNetworkMRWA_514_1.shp -o mrwa514.osm`

3. Create a JOSM style file that makes the names stand out:

way[cwy=Single] {
text: auto;
width: 3;
color: #c0c0c0;
text: "road_name";
text-position: line;
text-offset: 14;
font-size: 16;
font-color: lightyellow;
}

4. Mostly, road geometry is actually already in OSM, but the names and road types aren’t. To make it easier to find where to work, I apply filters in JOSM to hide everything apart from roads with no names. Then it’s just a matter of clicking a road, adding its info, and as soon as it’s got a name it disappears (thanks to the filter).

This is still pretty slow, but it’s safe. I’m sure there’s going to be lots of better ways to work with this, but at least some missing names (and geometries) are being fixed.

I’ll use source:name=Main Roads Western Australia to designate where the name has come from. The geometries I’m double-checking against Bing.