It’s been 18 months since Australia’s Geocoded National Address File (G-NAF) was made openly available by the Australian Government via data.gov.au. In that time, it’s been downloaded 3600 times, according to the Data and Digital Branch of the Department of the Prime Minister and Cabinet (the Department).
Recently, the Department sought feedback on G-NAF from end users and found 41% of respondents started using it only after it became openly available.
G-NAF was first released under licence by PSMA Australia Limited (PSMA) in 2004 after six years of research, a feasibility study, three pilots and two years of development. The aim was to create a comprehensive database of addressing knowledge that connected addresses recognised by government, with addresses adopted by the community, and a precise latitude and longitude.
G-NAF now underpins many policy and business systems that provide important services to Australians.
Recognising the value that can be derived from public data and its potential for promoting innovation, the Australian Government partnered with PSMA to make G-NAF publicly accessible at no cost to end users from February 2016. In feedback to the Department, 73% of respondents indicated they had achieved efficiencies or productivity growth as a result of open access to G-NAF.
What’s an address?
Address is a cultural construct, an artefact of human communication. It’s a generally accepted label for a location – a position on the Earth – but it’s transient and subjective. Governments gazette names or labels for suburbs and roads, and these generally reflect the language accepted by the community to describe place, but the community may use other language as well. People adopt unofficial addresses and use them consistently enough that they appear in government and commercial databases. Further, the collectors and custodians of addresses have inherent bias in their perception of what an address is, understandably skewed toward their core business purpose or need.
G-NAF contains more than 13.6 million principal addresses. It begins with PSMA collecting over 40 million addresses from Australia’s State, Territory and Commonwealth government agencies, testing the logical consistency of every address from every contributor and comparing the address components against other geospatial datasets it manages. The locality of the address is then confirmed as valid and the road name and road type are confirmed to exist within the locality.
PSMA subsequently assigns a confidence rating to each address as a measure of how widely the address is recognised. Once that occurs, a geocode – latitude and longitude – is assigned or verified.
In G-NAF’s May 2017 release, 96.18% of addresses were able to be geocoded at property-level, 3.57% at street-level and 0.25% at locality-level. It is the benchmark for geocoded address accuracy. Feedback to the Department indicates that 97% of those polled rate G-NAF data as ‘good’ to ‘very high’ quality.
G-NAF perceives an address to be a structured label for a place that can receive deliveries. You can go to the address, but G-NAF doesn’t tell you what you’ll find when you get there. It references a location, not a premises or building. A range of possible land uses might occur at an address.
What’s at an address?
PSMA recognised several years ago the need to answer the question of what’s at an address. Businesses and governments want to know about buildings and land cover at addresses across the country to enable better decision making and improved products and services. Hence, Geoscape® was conceived.
Geoscape combines advances in satellite imagery, machine learning and big data processing to create a digital representation of Australia’s built environment. It captures building footprints and heights, roof construction, surface cover, tree heights, the presence of solar panels and swimming pools and more. Geoscape uses G-NAF to link built environment attributes to a geocoded address. It’s the first time location-based information and data analytics capability have been combined and made available in this way on a national level.
Due to its vast scale and sophistication, the rollout of Geoscape is being staged. Release 3, in July 2017, saw the number of buildings captured by Geoscape double to more than 7.2 million. When coverage of the Australian continent is complete – anticipated to be by mid-2018 – Geoscape is expected to contain 20 million buildings mapped to geocoded addresses.
While G-NAF is derived from public data and is available under open licence, Geoscape is a commercial dataset (produced using commercial and public data) available through PSMA’s network of Value-Added Resellers.
Both G-NAF and Geoscape are underpinning policy and business systems that provide important services to Australians. This includes: navigation systems; infrastructure planning; business planning and analysis; logistics and service planning; insurance; telecommunications; and government service delivery and policy development, including emergency response and social services.
G-NAF provides the address and its position on the Earth; Geoscape describes what’s there.
Designing NextGen G-NAF
In the not-too-distant future, there will be no concept of spatial data as a unique data type. It will continue to be important as people inherently understand spatial as a way to categorise information, but it won’t be something considered in isolation.
Location will be a dimension of all data, a core component – because everything happens somewhere. As such, data users are beginning to look for location data that is structurally, as well as, financially accessible. They are looking to layer diverse data types into intelligent services to create knowledge and value, and they’re looking to do that efficiently.
Recognising that future, PSMA in consultation with the governments of Australia, is continuing to enhance and extend G-NAF. Creating the next generation of G-NAF involves a paradigm shift for national address management. It involves moving from the concept of a dataset to that of a knowledge base, one that stores complex structured and unstructured information and is connected to a broader digital ecosystem.
With the Australian Government making G-NAF available under open data terms and the resulting greater use of G-NAF across the economy, comes additional insights and feedback on the dataset. A common theme through that feedback is the need for G-NAF to be machine-readable using open data standards. Data users require the ability to ‘pull’ information they need to create knowledge, tools and value for their businesses and customers. This activity is increasingly being performed through machine-to-machine interactions.
To enable these transactions within geospatial supply chains, Program 3 of the Cooperative Research Centre for Spatial Information (CRCSI) is researching delivery of the next generation of spatial infrastructures. The program is leveraging semantic web technologies to iteratively build a spatial knowledge infrastructure test bed, starting with a small number of datasets, including PSMA’s G-NAF and Administrative Boundaries. It seeks to utilise linked data – involving standards, practices and tools for publishing and linking structured data on the Web – to publish data from both themes in an open standard, machine-readable format during 2018.
To maximise the benefits that come from machine-to-machine, transaction-based transfer of location information, PSMA will prototype continuous product maintenance processes for address data. Also, the unidirectional flow of content from traditional creators and custodians of address will ultimately give way to a closed-loop ecosystem, where users of the data contribute back to the source. A further key development will be improved feature-level metadata to provide greater insight into an address and its relevance to an organisation’s business need.
These fundamental enhancements to G-NAF will further reduce barriers to the use of location intelligence and help meet the changing information needs of the Australian economy, establishing an address reference that constantly aligns itself with user demand.
Jo Abhayaratna is Chief Technical Officer at PSMA Australia. Jo is also the new Chair of the Cooperative Research Centre for Spatial Information’s (CRCSI) Program 3 Board, Co-chair of the Open Geospatial Consortium’s (OGC) Smart Cities Domain Working Group and Chair of the OGC Australia and New Zealand Forum. Jo is focused on emerging data management and distribution practices to support PSMA’s continuing program of technology innovation.