Addresses are a critical part of our national data infrastructure, but address terminology can be really confusing. We thought it would be useful to explain how we talk about addresses in the same way as we described how we talk about registers.
What’s an address anyway?
It seems a simple question, until you consider the many variations of addresses. We know that how a user describes an address will depend on what they’re doing with that information. For example, the way they write an address on an envelope might be different to how they describe it if they’re giving someone directions.
Our working definition of an address at the moment is “a physical location that must be identified for regulation purposes including council tax and business rates, such as a house or office building.”
We’re deliberately keeping this definition flexible to reflect differing user needs. For example, if someone is trying to find a list of schools in one area, we can define addresses in this context as “a physical location within a school catchment area”.
When an address is created, it needs to be ‘recorded’ so that it’s recognised as a regulated location. We specifically use the term ‘recorded’ rather than ‘created’ as the address is added as a ‘record’ in a database.
In the context of addresses in digital services, we often talk about the need for a user to ‘enter’ an address. This is often when a user can type an address into a search service and select the right address from a suggested list.
Sometimes a user needs to find an address to get information related to that area, for example, find their closest school, hospital, or the council tax band for that area. We call that activity ‘finding a place’ and use the term ‘place’ to indicate a wider geographic area than an individual property or business.
Users write addresses in lots of different ways. For example, they may include the name of a house such as ‘Rose Cottage’ rather than its number, ‘72 Clover Lane’. Or they may make a typo or abbreviate a word like ‘road’ to ‘rd’.
A database will not contain all of the possible variations of an address. Instead, a ‘matching engine’ on top of a database of addresses can ‘match’ a user’s input to an address in the that database.
We use the term ‘normalisation’ to describe the process of turning the text typed by a user into a format that is compatible with the database and can be used to match against it. For example, searching for ‘72 Clover Lne’ with a typo will be 'normalised' to recognise that the user means 72 Clover Lane and display the appropriate results.
If a user is matched with an incorrect address, or finds a problem with a listed address, they should be able to tell someone so it can be fixed or improved.
The process of telling someone there’s a problem with an address and getting it fixed is known as a ‘feedback loop’. With millions of UK addresses, there are often errors that need fixing so feedback loops are extremely important.
An important thing about addresses is that different services request different formats of addresses. Some suppliers might just need enough information to deliver post. However, other suppliers might need more information about an address such as its latitude and longitude, political constituency, local authority or additional information linked to an addressable location.
We use the term ‘presenting an address’ to describe the process for taking information from a database and turning it into a variety of usable formats.
We describe any physical location where a person, organisation or the like is located or may be reached as an ‘addressable location’. One whole address can be made up of lots of addressable locations.
One addressable location could be the first floor of an apartment block. Another addressable location could be the name of the apartment block itself.
Testing this language
We’re starting to test this language even more with our users to make sure we’re using terminology that everyone understands. This list is by no means complete and we have lots more detailed terminology that we haven’t explained here.
We’ve also shared our choice of language and definitions with a community of government organisations that use addresses so we can start a conversation about address terminology. If you’re working in government and would like to get involved in the community, send us an email.
You can follow Jen on Twitter, sign up now for email updates from this blog, or subscribe to the feed.
Comment by Wolfgang (Schindler) posted on
It's interesting to see that once you submit seemingly trivial aspects of everyday life to a closer scrutiny, you will face the challenge to create an adequate model of this section of reality which is far from trivial, but very fascinating.
Comment by Digital developer posted on
Agree with Wolfgang. Sounds simple yet complicated. What is "the single version of truth" for addresses in UK? Is it ordinance surrey's data? Even that is inconsistent and the format changed slightly recently.
Comment by Mick Phythian posted on
How does this all fit in with:
Based on BS7666 these are a series of address-based datasets bringing together a variety of local datasets and perfecting them at the national level. This has chiefly been organized on a profit-making basis by Information House, who negotiated a suite of datasets for Local Authorities under the same Mapping Services Agreement (MSA) that the OS is now part of (see above). These N-Initiative datasets, many of which are part of the MSA, include:
The National Land and Property Gazetteer (NLPG) : http://www.nlpg.org.uk/
An aggregation of Local Land and Property Gazetteers (LLPG), tax details and electoral registers constructed by Local Authorities, this will ultimately provide an unambiguous address-point identification of land and property. The advantages in this for tying together previously disparate government activities (various taxations, for example) are great, as are the commercial opportunities – profits from which will be fed back to Local Authorities via Information House (Baker, 2003). It will also enhance public interactions with government datasets via its use in web-based searches. The development, initiated in 2001, is now contracted to Property Intelligence plc’s Intelligent Addressing consultancy unit as part of the Mapping Services Agreement.
The National Street Gazetteer (NSG) : http://www.thensg.org.uk/
This collates Local Street Gazetteers (LSGs) and Associated Street Data (ASD) prepared by local highway authorities into a national level identity reference for streets. As this is a relatively mature dataset Information House can use the resultant profits to fund the development of the other datasets. The continuing work is currently contracted to Intelligent Addressing as with the NLPG.
The National Land Information Service (NLIS) : http://www.nlis.org.uk/
Not so much a coherent dataset as a version of GIgateway for land and property. This website negotiates electronic searches across multiple government and private sources of land and property information, at the moment to principally support the conveyancing process. Set up in 1998 by HM Land Registry and now managed by Information House with inputs from organisations including the OS. Built, in part, on the NLPG dataset.
National Land Use Database (NLUD) : http://www.nlud.org.uk/
This project aims initially to locate all potential brownfield (Previously-Developed Land : “PDL”) sites in the UK, then extend the survey to cover all land uses. It is a partnership of the OS, English Partnerships and the ODPM. The NLUD-PDL survey is advanced, while the longer term project has currently reached the point of generating a “Land Use and Land Cover Classification” but is under consideration for further development.
Co-ordinated Online Register of Electors (CORE)
Building on a failed IDeA project “LASER” (Local Authority Secure Electoral Register), CORE aims to provide a address-point list of all registered voters in the country at the national level, maintained by Local Authorities. This would enhance the potential for the cross-comparisons of datasets and easier management, and, more notably from the government’s point of view, commercial exploitation. This would be tied with the NLPG into a “One stop change of address” system, nicely put across as for the convenience of those moving around the county, but, of course, equally about tracking the citizenry.
Comment by Jen Lambourne posted on
Hi Mick, thanks for your comment.
The terminology we’ve mentioned here isn’t intended to disagree with or replace any of the terminology that are already in widespread use - especially when they are underpinned by established standards such as BS7666.
What this blog post highlights are the working definitions that we’ve been playing around with in a community of government organisations that use addresses. We’re trying to develop a shared understanding of some common activities regarding the use of addresses and use the same language when it comes to describing them.
Comment by Dominic posted on
One recent problem I have experienced is the GOV.UK digital sites not recognising new UK postal addresses. How often do you update the Postcode Address File from Royal Mail?
Comment by Jen Lambourne posted on
Sorry to hear you've had trouble with new postal addresses. The GOV.UK team have been doing some work to improve this using the latest Office for National Statistics postcode dataset, which is updated each quarter. The GOV.UK team wrote about their work in this blog post, which might be of interest: https://insidegovuk.blog.gov.uk/2016/06/27/improving-the-postcode-matching-on-gov-uk/