GSS Logo

High quality statistics, analysis and advice to help Britain make better decisions.
 

The new ONS National Statistics Address Lookup (‘NSAL’) is nothing to ‘sniff’ at…..

Posted by
|

Street names and addresses are important. Very important. There was a time when they were mainly about helping the postman deliver your letters but these days they do everything from helping your weekly shop reach the right house, to helping your sat nav get you where you need to go, to helping work out what your insurance premiums will be, to helping the emergency services save lives.

It has been known for some time that addresses on their own are limited not only by their ability to repeat themselves (does every town have a London Road?) but also the ability for a road to have more than one name or in one case in Seattle for two roads with the same name to meet despite having different numbering systems. Organisations have therefore sought increasingly complex ways to work out which ‘No 3 London Road’ a person might be referring to or even whether Flat 1, 3 London Road is the same as Flat A, 3 London Road or 3A London Road.

These increasing attempts to uniquely identify a property eventually led us to the introduction of the National Land and Property Gazetteer’s Unique Property Reference Numbers (UPRNs) that assign a unique code to pretty much anything that you might want to uniquely identify. Pretty much every property in the country has a UPRN, from mansions to beach huts and ponds to life belt buoys (yes really!). As the name suggests, this number uniquely identifies a property, so if I’m talking about UPRN 10033544614 I’m talking about Buckingham Palace, London SW1A 1AA.

London map

So why do the UPRNs matter for statistics? Well firstly they are an important part of the statistical process as the GSS Geography Policy lists addresses as the best level for geocoding statistical observations and allowing them to be recast to different geographies. Secondly they are a good linking mechanism for bringing different datasets together and consolidating them to produce statistics.

The problem has been that the UPRNs were proprietary data owned by GeoPlace (the organisation set up to oversee the amalgamation of the National Land and Property Gazetteer and OS AddressLayer 2) and so could only be used in a limited way for data sharing, particularly where non-government organisations were concerned.

All this changed a few months ago though, when Ordnance Survey allowed AddressBase® users to release UPRNs as open data (after a request by ONS). This move allows organisations to start using the UPRN as a ‘golden thread’ – a link between different datasets that allows data to be consolidated and consumed at the address level in an open way.

This is particularly important for the 2021 Census as it looks at moving away from traditional Census data capture to an increasing reliance on administrative and big data. We expect both the traditional Census enumeration and the development of the administrative data outputs to be done on the basis of the open UPRN.

We are also aware that the requirement for linking addresses (through the UPRN) to statistical geographies is much wider than Census, ONS or even the GSS. The opening up of the UPRN offers the opportunity for all producers and users of statistical data to add value by linking address level information to the geographic hierarchy in the same way that we currently do for postcodes with the National Statistics Postcode Lookup.

On that basis ONS Geography have recently released the National Statistics Address Lookup (NSAL for short but seems to be pronounced nasal within ONS) that relates the UPRN for each GB address to a range of current statutory administrative, electoral, health and other statistical geographies via ‘best-fit’ allocation from 2011 Census output areas (OA).

The NSAL will be issued every 12 weeks and is designed to complement AddressBase®. This new product supports the GSS Geography Policy to reference source data at the lowest possible geographical level using a standard identifier – for addresses this is the UPRN. When the data being referenced relates directly to a real world object (for example a building or section of street), they are linked to these objects by the UPRN.

With the official launch of the NSAL product last week we are now considering whether there is enough user demand for producing an ONS Address Directory (ONSAD). Like NSAL, this product would also have GB coverage and list UPRNs against a number of standard geographies, such as Local Authority District, Ward, Parliamentary Constituencies etc but by using a direct point-in-polygon methodology rather than ‘best-fit’. A very small proportion of URPN allocations will therefore differ between NSAL and ONSAD. So over to you – please let us know if this is something that you would be interested in.

The NSAL product, like all our geography products, can be downloaded from the Open Geography Portal via the ‘Download products’ tab in the ‘Address Products’ section. For further information/help/guidance on all our Geography products and services please contact ONS Geography.

Comments

  1. Profile photo of Howard Askew

    Howard Askew

    February 26, 2016

    “A very small proportion of URPN allocations will therefore differ between NSAL and ONSAD. So over to you – please let us know if this is something that you would be interested in.”

    Hi, I think I’d rather have a firmer steer from the statistical centre about which should approach should be adopted – point-in-polygon or best-fit. If the golden thread is being used inconsistently, surely this erodes the advantages,

    How does best fit work at an address level? There’s just one set of coordinates – which could be the front door of a property, the centroid of a plot of land, or any point within a boundary. How is that allocated to an OA by best fit?

  2. Profile photo of ons.geography

    ons.geography

    February 29, 2016

    Hi Howard, Point-in-polygon is used to allocate the UPRN to the OA in which it actually falls. The resulting OAs are then used in the best fitting to other geographies. Hope this helps.

    Barb Lennards – 2016/02/26 at 11:32 am

Add a comment

You must be logged in to comment.