Open Addresses and Linked Data
There has been a lot of talk recent about the need for an open national address dataset. As someone who has campaigned tirelessly for open postcodes, and, in my previous role at a local authority, who is only too aware of the issues surrounding licensing of geodata, this is something I’m passionate about too.
However, all the talk recently has been around making the Postcode Address File (PAF) public, which, given that the Royal Mail is about to be privatised, I can understand.
That said, after a discussion on the My Society mailing list, I’ve heard directly from the people who have used the PAF and their experiences haven’t been good. The data quality is inconsistent and it’s not maintained very well. I suspect. If this data is opened up, there may not be an argument for a privatised Royal Mail to continue maintaining it.
There is, however, a third way. The UK already has a super-accurate, up-to-date and well maintained database of address data. It’s managed by local authorities as part of their day-to-day activities, and regularly updated.
That database is the National Land and Property Gazeteer (NLPG). It is managed centrally by GeoPlace, a public sector partnership between the Local Government Association and Ordnance Survey. When local authorities collect data about a new property, they enter it into their databases in an agreed standard. This local data is then submitted to GeoPlace and entered into the national gazeteer.
Crucially, every property in this database is given a unique idenfier, known as a Unique Property Reference Number (UPRN), which, if this data were opened up, gives us great scope for linking datasets together and giving every single property its own URI.
This would have the benefit that when two datasets are talking about (for example) 123 High Street, Anytown, we’re definitely talking about the same property.
If we could open up this data, not only would we have a reliable, well-maintained source of address data, we would also have a key piece in the jigsaw of getting truly linked geographical data.