I’m currently studying an MRes at University College London and have a topic i’m interesting in covering. To test data availability i used the overpass script using the code below and was surprised on how few sites it threw out.
I ran it over my local neighbourhood and noticed there were many sites not picked up, i know the sites are ‘undeveloped’ through visual inspection and famialirty. Looking on OSM with the map data layer turned on these areas appear not to be classified. I’d like to get a quantem of how much land within a certain area is undeveloped and as a work round thinking i could do this by exclusion of the recognised tags summing the out put areas and subtracting this from the actual physical area. Before going down that route i want to check i’m not missing something much more simple! Any and all help gratefully appreciaed.
It depends on how users understand and use the ‘brownfield’ tag.
I ran the same query over a part of London I’m familiar with. It missed some extensive brownfield sites but picked up three small ones in a town centre. The extensive sites are actually green to the eye and from the air because they get little maintenance and vegetation has been allowed to grow. They are former industrial land, now fenced off spaces, awaiting redevelopment. They are currently tagged as ‘natural = scrub’, which is correct in one way. It may not occur to most mappers that this land is technically ‘brownfield’.
The correct tagging is probably both the scrub and the brownfield tags used together, but this should be discussed and recorded on the wiki.
What I’d suggest that you do is look at the areas in OSM that you’re interested in and see what they are mapped as (if anything). If you zoom in and use the “question mark” at the right, you can query what things are, so here
you can find that http://www.openstreetmap.org/way/163386158 is “landuse=brownfield” with a description of “Former factory, awaiting redevelopment”. However, quite a lot of landuse isn’t mapped (yet) - just compare the brown farmland on http://www.openstreetmap.org/#map=10/52.7517/-1.1501 with the actual amount of farmland in that area, and you can see that there is lots still to do.