[Talk-GB] Import UK postcode data?

[Russ Phillips]

OsmAnd is what made me think of it. I'm using that for directions a lot 
now, and it's common these days to give a postcode to plug into a sat 
nav, but they frequently don't work in OsmAnd.

I was somewhat surprised that it hadn't already been done, and I accept 
all the reasons for not doing it.

I'll look into alternatives.

Russ


On 02/10/2019 15:52, SK53 wrote:
> I really see no point. The data are already present in Nominatim  
> (albeit perhaps not up-to-date) and search is the ONLY thing that 
> so-called postcode centroids can help with. DE24 (Sinfin) was imported 
> long ago, see this overpass query <http://overpass-turbo.eu/s/MNZ>.
>
> What can be done is attempt to assign postcodes to streets, and 
> various incomplete efforts have been made over the years. Open Data 
> provides full addresses for around 70% of UK postcodes (principally 
> Companies House, Food Hygiene & National Register of Social Housing).
>
> What would be useful is a maintained set of postcode information based 
> on codepoint open/ONS postcode data/OS Local. The sorts of things 
> which it would be useful to know are:
>
>   * Is the postcode centroid co-located with others (e.g., delivery
>     offices, some businesses, blocks of flats)
>   * Can the postcode be unambiguously assigned to a street & post town?
>   * Is the post code in use or not (ONS is now obviously 8 years old,
>     but still potentially useful). Greg's FHRS tracker does appear to
>     indicate a degree of churn with inner-city postcodes (although
>     some of this will be inadvertent use of more general rather than
>     specific postcodes by people filling in the FHRS forms)
>   * Is a postcode the sole postcode for that street?
>
> Somewhere I have an old CPO table with some of this data populated. I 
> think Geolytix had summary info associated with their post code sector 
> shape files as well.
>
> Adding addr:postcode to streets which have a single postcode is in my 
> book fine: numerous LAs put this on street signs (Rushcliffe & Gedling 
> for sure); it's an intermediate step to adding the addresses to 
> houses  Once the postcodes which obviously belong to a single street 
> are eliminated it's often easier to work out where the others belong.
>
> Note that other than for FHRS we have no good source for Northern 
> Ireland postcodes at all. Equally assignment of rural postcodes is 
> quite a bit harder than urban ones. In re-reading bits of Chris's blog 
> last night I came across a post of his 
> <https://chris-osm.blogspot.com/2017/03/have-you-moved.html> showing 
> that some postcodes move huge distances between releases.
>
> Jerry
>
> PS. The Sinfin postcodes possibly should be removed as they were added 
> IIRC before Mike Collinson's discussions with OSGB about OS Open Data.
>
> On Wed, 2 Oct 2019 at 13:44, Russ Phillips via Talk-GB 
> <talk-gb at openstreetmap.org <mailto:talk-gb at openstreetmap.org>> wrote:
>
>     Hi,
>
>     I'm wondering if it would be feasible and advisable to import the
>     UK postcode data from OS OpenData Codepoint
>     <https://wiki.openstreetmap.org/wiki/Ordnance_Survey_OpenData#Code-Point_Open>.
>
>     The licence is OSM compatible. My thinking was that we could
>     create a node for each data point and set the addr:postcode tag.
>     This would be useful for routing software like OsmAnd, since it
>     would allow a user to enter a postcode as a destination.
>
>     I'm happy to do the work, but the import guidelines
>     <https://wiki.openstreetmap.org/wiki/Import/Guidelines> say that
>     imports should be discussed on the imports@ list and the
>     appropriate local communities, hence this email.
>
>     Russ Phillips
>
>
>     _______________________________________________
>     Talk-GB mailing list
>     Talk-GB at openstreetmap.org <mailto:Talk-GB at openstreetmap.org>
>     https://lists.openstreetmap.org/listinfo/talk-gb
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.openstreetmap.org/pipermail/talk-gb/attachments/20191002/e1e9abcf/attachment.html>