[Talk-GB] UK Postcodes
sk53.osm at gmail.com
Mon Sep 26 13:29:12 UTC 2016
I just re-read a post
I wrote nearly 3 years ago. I think a lot of it holds true today, so I've
copied the main points here :
1. The simplest, but not necessarily the easiest target, is to map at
least one postcode in each postcode sector. This is harder than it appears
because obvious things to map in sparsely populated rural areas may require
surveys. For instance FHRS data has two B&Bs in Port Wemyss on Islay, but
the names are not shown on the OS Open Data StreetView. Similarly a degree
of caution must be exercised on farms in the Rhinns of Islay and on the Oa
because individual farmsteads may include two or three properties (perhaps
all owned by the same extended family, but nonetheless distinct.
2. Achieve 5% completion. This reflects a DOUBLING of current postcode
data, and therefore must be regarded as ambitious. This is however, the
minimum condition for breaking the back of the postcode problem. I believe
with a concerted effort we could achieve this in 3 months, using
conventional crowd-sourcing techniques.
3. Achieve 10% completion. A second doubling will probably require more
tool based support. The obvious targets are semi-automated matching of FHRS
& Land Registry data, and semi-automated identification of single postcode
4. Postcodes along major roads (A & B roads). These may require some
survey work, but again because many retail outlets are along such roads
there is already a decent amount of information available from FHRS.
This was December 2013, so perhaps 5% and 10% should be nearer 10% and 20%.
I don't have up-to-date figures but back in May 2015 we had 73,372 full
well-formed postcodes for GB (not whole of UK) which is still under 5%.
These were located in just under 8000 postcode sectors (out of a total of
12,300 or so, with another 1000 populated in the last year). FHRS data has
information on nearly 250k postcodes (inc NI) and 10k distinct postcode
sectors. All these figures are based on raw strings, i.e., not checked if
valid or in the right place. We still have thousands of schools mapped
without postcode (even some where ref_edubase was added) so this is another
fairly easy target.
The big difference from 3 years ago is that we have more people interested
in creating tools to assist these processes: something where the 3 month
timescale is better than a shorter one.
We have needed to get more address data for some, but on its own it's not a
very strong motivator. My hopes for making big progress with Land Registry
data were dashed once OpenAddresses and Owen Boswara clarified the 3rd
party content in the data, and similarly the OpenAddresses project finished
without having much in the way of additional data to offer us. (I still
believe that there's scope in their approach and they built some
interesting tools, but it was predicated on already having a decent amount
of usable open data). When one looks at the formidable success of BANO in
France there must be scope for something similar in the UK.
I'm going to try & update my PC completion maps for the UK. I have some now
but I know I have lost data from filtering the gb file.
On 26 September 2016 at 11:44, Brian Prangle <bprangle at gmail.com> wrote:
> It looks like the next UK Quarterly Project will be based on improving
> address data for town centres using the food hygiene dataset. Why don't we
> have a push generally on postcodes too, not limiting it to town centres?
> On 26 September 2016 at 11:25, David Woolley <forums at david-woolley.me.uk>
>> On 26/09/16 10:19, Owen Boswarva wrote:
>>> That could be done but it's not straightforward; you'll get a lot of
>>> overlapping postcode sectors and sectors with non-contiguous parts.
>>> GeoLytix produced an open dataset like that some time ago:
>> In my view, inferring polygons is something that should only be done in
>> the data consumer, as they involve creating data that cannot be justified
>> from the input data.
>>> On 26 September 2016 at 09:39, Colin Smale <colin.smale at xs4all.nl
>>> <mailto:colin.smale at xs4all.nl>> wrote:
>>> How about deriving polygons for the postcode sector level (XX9 9)
>>> from the centroid point cloud, and adding the polygons to OSM? I
>>> don't know how many that would give, but it would be a whole lot
>>> less than 500k and still at a very usable level.
>> Talk-GB mailing list
>> Talk-GB at openstreetmap.org
> Talk-GB mailing list
> Talk-GB at openstreetmap.org
-------------- next part --------------
An HTML attachment was scrubbed...
More information about the Talk-GB