GeoPlace is creating the National Address Gazetteer on the twin foundations of the National Land and Property Gazetteer (NLPG), created and maintained with input from all local authorities in England and Wales, and Address Layer 2 (AL2) which is maintained by the Ordnance Survey and incorporates Royal Mail PAF data.
Both datasets contain grid co-ordinates, property classifications and references to external data sets such as the Valuation Office Agency Council Tax and Non Domestic Rates data, together with links to their respective street gazetteers, the Integrated Transport Network (ITN) and the National Street Gazetteer (NSG).
The creation process
The integration of the source data was overseen by Intelligent Addressing using a combination of in-house proprietary software, experienced data technicians and OS spatial skills to match and review millions of records and more importantly, co-ordinate the efforts of 348 local authority gazetteer custodians to resolve anomalies using their local knowledge and unrivalled experience and access to local data.
The initial stages of the process produced immediate results with some 84.5% of the PAF sourced AL2 data having an exact match in the NLPG. Subsequent application of customised matching algorithms to allow for minor differences in building address, street name, locality and town name, organisation name, property classifications or grid co-ordinates and a review of common links to third party data, identified a further 12% of records that could be safely synchronised.
Taking the most conservative approach possible, the residue from the central creation process of some 380,000 records classified as "dwellings" has been sent to local authority gazetteer experts for resolution, with commercial records to follow. Independent validation of the proposed matches has been undertaken by the OS address management team.
Although the Royal Mail sourced multiple residency data from AL2 has been reviewed, current thinking is that the local authority sourced NLPG data has a wider variety and better coverage of such "child" records and that the national address gazetteer will not include this data or reference to it.
The process of synchronising the OWPAs from both AL2 and the NLPG is complicated by the different sources used. There are AL2 classifications such as shafts, parks, hoppers etc. that do not appear in the NLPG, and NLPG classifications such as car parking spaces and advertising rights that do not appear in AL2. The process is further complicated by the classifications such as "electricity sub-station" that appear in both but against different criteria, the use of addresses based on descriptive locations rather than formal names, some non specific grid co-ordinates and the use of descriptive street names. However, these records are of vital importance to existing NLPG and AL2 users and, with input from the local authorities and help from site surveys by OS cartographers, the intention is that all will be synchronised and included in the national address gazetteer.
The above processes are designed to allocate UPRNs to all of the AL2 records and incorporate them, along with their NLPG counterparts, into the national address gazetteer. There they will be joined by a significant number of records sourced purely from the larger NLPG, for example there are currently estimated to be 1.8m NLPG sourced residential records. Each of these records will come with some form of corroborating evidence to justify its inclusion in the national address gazetteer, for example a link to Council Tax, Electoral Registration or a planning permission.
The final product
Even as the process of synchronisation nears its final stages, the business rules are being formulated to deal with multiple inputs from different sources within different time frames that will be at the heart of the unique national address gazetteer maintenance process. In addition, rules to resolve the outstanding anomalies in grid co-ordinates, classifications and minor address or organisation name differences that arise from the data synchronisation are being tested.