Comments in-line On Tue, 5 Aug 2025, 11:32 Andy Mabbett, <[email protected]> wrote:
> In a series of blog posts, the latest of which is: > > https://www.owenboswarva.com/blog/post-addr12.htm > > my friend Owen Boswarva describes how he has obtained council tax > bands, addresses and UPRNs under OGL, from Leeds, Birmingham, > Lichfield and Hackney councils (and some of the data from Nottingham) > and made them available. > > Can we now import the data, using address matching? I suspect, unless rigorously tested, this will be error prone. For a single council, even those as big as Birmingham, I'd expect a concerted effort to hand match will be as fast. Can we use it to > find buildings that are mapped as single dwellings, but which are in > fact subdivided? > UPRNs can be used to do this directly, just do a point in polygon query comparing UPRNs and buildings extracted from OSM. (You can also do this OS Local buildings which is a good, but not perfect, way to reduce noise in the UPRN dataset). The fly in the ointment is that historical UPRNs are included too, but I suspect the majority of these will be multi-occupation buildings. FWIW I assigned unambiguous UPRNs to, IIRC, around 35k Nottingham addresses https://github.com/SK53/osm_uprn/blob/master/ng_osm_uprn_lu.csv. I've remarked before, but, at least in England and Wales, UPRNs are often ordered by housenumber. This often allows additional addresses to be inferred. I have an imperfect algorithm to generate housenumbers from UPRNs. The other file in that github directory is data from an intermediate step. Jerry
_______________________________________________ Talk-GB mailing list [email protected] https://lists.openstreetmap.org/listinfo/talk-gb
