Address information is integral to databases ranging from the simplest contact list to the most sophisticated enterprise systems. An address is easily created, collected and applied. But being based in language, the very features that make addresses accessible and easy to use also mean that they can be ambiguous, open to interpretation, and lacking in the mathematical consistency required for rigorous matching and location analysis.
Australia’s authoritative national address reference file for addressing the nation is G-NAF. The Geocoded National Address File (G-NAF) was developed by PSMA Australia, the company established by the governments of Australia to provide an independent and self-funded means for collaborating on national geospatial matters.
“Since its introduction in 2004, G-NAF has developed a strong track record of supporting business critical systems in a wide range of industries including the telecommunications, insurance, banking and finance and utilities sectors,” said Dan Paull, chief executive officer of PSMA Australia. “G-NAF also underpins Australia’s National Address Management Framework that has been adopted by governments at all levels.”
In early 2012, PSMA Australia entered into an agreement under which NBN Co will use G-NAF for address information to support the planning, roll-out and operation of the national broadband network.
G-NAF brings together datasets supplied by the state and territory governments of Australia, the Australian Electoral Commission and Australia Post.
“As these are authoritative and relatively independently assembled datasets, each with its own strengths and weaknesses, G-NAF merges the supplied addresses into a single G-NAF address record with feature level metadata documenting its lineage & quality. Address confidence is established by the number of occurrences in the different datasets,” said Dan.
This analysis and comparison is conducted within a geospatial environment so that each component of each address can be matched against the official geospatial region to which it relates: state, locality, street and finally property.
“The production process creates extensive information detailing the variations that exist between the official principal address and the address that people use,” said Dan. “G-NAF features extensive address level alias listings and street and locality aliases, including vanity names. Alias addresses are addresses, other than the principal address, that refer to the same physical location as another address record. G-NAF is a hybrid dataset combining the strengths of each input dataset (both address and spatial) to deliver a premium quality reference set.”
Figure 1: Example of how G-NAF manages vanity addresses at a locality level to avoid duplications
PSMA Australia works closely with the data contributors to ensure the dataset supports the emerging needs of Australia’s geospatial data user community. PSMA Australia implements rules to match alias/unofficial information to the official record as part of the data processing. G-NAF’s founding principal, however, is to respect the integrity of the addresses supplied by the authoritative contributors.
“Many of the issues resulting in missing or duplicated data are well understood and documented,” said Dan. “Multi-dwelling sites and private estates such as caravan parks, aged care facilities and gated communities are a common source of address variation. The state and territory governments are particularly focused on a range of activities to better capture the addresses for these types of properties.
“Ranged addresses cause an estimated 300,000 duplications within G-NAF. A ranged address such as 2-8 Smith Street may be recorded within different data sources as 2 Smith Street, 2-6 Smith Street or 2-10 Smith Street.
Figure 3: Example of ranged addresses
“More than 90% of these duplications will be resolved in the August 2012 release of G-NAF, where PSMA Australia will adopt the address supplied by the jurisdiction as the principal address and create aliases for the alternatives.
“The use of different number systems (Unit 1 or 1A) for multi-dwelling sites is a more difficult issue to resolve and PSMA Australia is working closely with address contributors with a number of management strategies under consideration. “
Figure 3: Example of how G-NAF manages a common source of address duplication at a street level
Dan said that despite G-NAF being the most current address reference, data currency has always been a challenge.
“The data collection and processing takes time and this means that there is latency in the quarterly release,” said Dan. “In late 2012, PSMA Australia plans to introduce G-NAF Live, a near live database of the most recent authoritative addresses provided by the state and territory address custodians.
“PSMA Australia receives address data as frequently as the state and territory custodians can supply it,” said Dan. “The address data is then converted into a single data model (using the NAMF & AS4590 compliant model), loaded into the live database and exposed to address verification through PSMA Australia’s award-winning technology, PSMA Systems.”
“Designed to support the National Address Management Framework, G-NAF Live and supporting services will generate substantial improvements to address currency and accuracy. G-NAF Live will be a key component in a feedback loop for investigating and correcting addresses as well as near real-time access to new and corrected addresses.
“This capability of delivering changes in the national reference dataset as they occur within the state and territory custodians represents a significant advancement for national address management and will set a new benchmark, not just in Australia but around the world.”