Open Addresses Discovery Phase

Jeni Tennison

Geospatial data underpins practically everything, because everything happens somewhere. A lot of things happen somewhere with an address: our schools, hospitals, museums, supermarkets, places of work, homes. So address data — data about what addresses exist and where they are — is an essential underpinning for our national information infrastructure.

The use and utility of open data in the UK has been limited by the fact that a lot of geospatial information, and in particular data about addresses, is closed. It is only available to restricted sets of people (mostly those who can pay) and for restricted purposes. So it can’t be used freely to create new, innovative products and services or simply to help everyone work more effectively by using the right addresses.

None of the attempts to unlock existing sources of address data in the UK have been successful. The Postcode Address File (PAF) is maintained by the Royal Mail, and while it is feasible for a private company to open up data in the public interest, if Royal Mail were going to do so, it probably would have done so by now. The National Land and Property Gazetteer, available within the AddressBase suite of products from Ordnance Survey, contains data from the PAF and therefore cannot be opened while PAF remains closed.

The recent Katalysis report, commissioned by BIS, on an Open National Address Gazetteer says (my emphasis):

UK society relies heavily on address data and current products have helped greatly to create benefit. The review has determined that Open usage would result in substantial and valuable growth among new user types and with even greater community benefit. The recommendation is that a basic address product should be free to all users at the point of use.

It’s time for us to build an Open Address Database together.

The success of OpenStreetMap has demonstrated what a community can do when it works together to create a common good. We need to replicate that success around what is, in some ways, a much smaller problem: a database of addresses. As Steven Feldman said in his blog post Enough of this AB PAF, I need an OAF:

If we can’t make PAF open, surely we have the skills and raw data to start building an Open Address File? If we worked smart with the Open Data that is available, added in what we already have in OpenStreetMap and did some smart conflation and interpolation we would be on our way to creating an Open Address File. It would be patchy to start with but once we had a framework we could find ways to fill in the gaps.

And the great thing is that the Open Data User Group and the Public Sector Transparency Board, both run out of the Cabinet Office, agree, to the extent that they have funded ODI to, in the first instance, look at the feasibility of developing an Open Address Database.

The initial Discovery phase, which is all that’s been funded thus far, and will probably run through the summer, has three general aims:

  1. To understand the data that’s available, its coverage, and how the database can be structured such that we can trace back its origin to open sources or individual entry to demonstrate the data hasn’t come from PAF and therefore isn’t subject to Royal Mail’s IP rights.
  2. To scope out and understand the governance framework and business models that could be adopted by such an enterprise, to ensure that it’s sustainable in the long term.
  3. To engage potential users, providers and developers to understand their desires and requirements, so that whatever is built is useful.

We’d love to know what you think. You can help right now by:

  • commenting, blogging or tweeting your ideas (use the hashtag #openaddresses) about what an Open Address Database could and should be, and how you would use it
  • adding open datasets that contain addresses to the list on our Wiki
  • getting in touch with to tell us how you’d like to be involved