Many of you may have seen a tweet <https://twitter.com/owenboswarva/status/1279853518476775427> from Owen Boswarva last night. It's prompted me to do what I planned to do anyway.
Owen looked at West Bridgford, a suburb/town within Greater Nottingham, and found 77% of INSPIRE polygons could be directly associated with a single UPRN. West Bridgford is a wealthy area and predominantly owner-occupied houses, although there are some blocks of flats. I've repeated this for Nottingham which has much more social housing. There are in total 102.5k polygons for Nottingham, of which 77k contain only a single UPRN. I then linked those to UPRNs I'd already linked to OSM buildings and joined the two datasets. I retained 26.5k polygons, or about a third of the candidates. I tweeted <https://twitter.com/SK53onOSM/status/1280258482315448320> some images. I've only used addresses on building polygons: a significant proportion of addresses are still mapped, or will continue to be mapped, as nodes. Extending the matching to use these would, I anticipate increase the matched proportion to 50% or slightly better. In some areas improving alignment of OSM data would also make a difference. One of the main things I notice is that it highlights fairly simple improvements one can make in OSM data. Polygon coverage is very high: the things which I've noticed missing are : churches, some public open space, allotments (of great antiquity), some schools, almshouses, railway land, and Army Reserve Centres. At least one set of allotments has separate INSPIRE polygons for each plot. Large polygons in residential areas tend to be leasehold or social housing landlords. Jerry
_______________________________________________ Talk-GB mailing list Talk-GB@openstreetmap.org https://lists.openstreetmap.org/listinfo/talk-gb