Dan Paull, Chief Executive Officer, PSMA Australia
This paper was presented at EARTH: FIRE AND RAIN, Australian & New Zealand Disaster and Emergency Management Conference, Brisbane, 18 April 2012
The daily interactions between government, the private sector and the community result in thousands of new addresses being created and captured in Australia every week. This address information is collected and stored in many different ways, creating the potential for addresses to vary widely in quality and accuracy. Disaster and emergency management agencies have long recognised that imprecise address information has the potential to place lives at risk through delayed emergency response.
In recognition of the economic and social benefits associated with improving address quality, the governments of Australia have adopted a national coordinated approach to address management. The National Address Management Framework (NAMF) is a practical, consistent, standards-based framework to guide the process for verifying and exchanging address information. The NAMF goal of ‘one address = one location’ is supported through the use of G-NAF (Geocoded National Address File), Australia’s authoritative index of physical Australian addresses and their location.
Developed and managed by PSMA Australia, G-NAF has been the leading national address dataset since 2004. The dataset is updated on a quarterly basis. Not surprisingly, there has been an increasing demand for more frequent updates, particularly from the disaster and emergency management sector.
In the last few years, PSMA Australia has been strongly focused on working though the challenges associated with delivering this dataset more frequently. This will be achieved with the release of G-NAF Live, a new service that will provide near real-time access to new addresses gazetted by state and territory governments.
This paper will discuss the benefits of real-time address management for disaster and emergency services agencies and the efficiencies generated by access to authoritative national datasets.
Access to accurate addresses is a risk management fundamental for effective disaster and emergency responses – whether it is for the rapid despatch of emergency crews or disaster planning and modelling activities. Disaster managers and emergency service providers require access to the highest standards of address accuracy and coverage, as well as up-to-date information. Most services invest in considerable resources for the establishment and maintenance of their address holdings.
The need for this effort reflects the incredible variations in addresses that have typically arisen and have been caused by a wide range of reasons that tend to reflect the different requirements and perspectives of the address creators, collectors and users. There is also often a significant divergence between the official address and the one that people use. This is particularly highlighted in emergency situations where stress, confusion or an unfamiliar location cause people to provide inaccurate address information or to rely on other landmarks and features to identify their location.
This paper provides some background and discusses the complexities around addressing in Australia and current activities that will deliver significantly better quality address management support for disaster and emergency management. This includes exciting developments associated with the implementation of the National Address Management Framework (NAMF) that offer the disaster and emergency management sector considerable potential to eliminate duplications of effort to achieve cost and resource savings without compromising operational efficiency or public safety.
To provide some background to the reasons behind address issues, the starting point is to look at the major sources of official address information in Australia. It should be noted that considerable investment and commitment is required to develop and maintain a comprehensive address dataset. To warrant this, the dataset must serve the critical addressing objectives of the organisation it is supporting. Taking this a step further, it is not unreasonable to note that no single entity currently has an interest in making use of every address in Australia, although this may change in the future. This means that even very comprehensive datasets have some limitations when it comes to their coverage.
Some of the best address datasets for Australia include those provided by the state and territory governments, the Australian Electoral Commission and Australia Post. Each of these cover around 80% of addresses in Australia.
State and territory governments
The state and territory governments collect addresses to support land and property management. This information is largely provided to the state government by local government. These address datasets offer excellent coverage as well as access to up-to-date address information. Increasingly these datasets are also including addresses for proposed developments, providing advance notice of impending changes. Some of the limitations of these address datasets occur in relation to their coverage of private estates and complex addresses. With addresses being captured at the point of development, these addresses may not actually reflect how the addresses have evolved during the site’s occupancy.
Australian Electoral Commission
The AEC collects addresses to support its electoral roll management responsibilities. The AEC dedicates considerable resources to ensuring their address their dataset is accurate, particularly in the lead up to elections. However, this address dataset’s primary focus is on residential addresses rather than commercial or industrial addresses.
Australia Post’s address database is structured to support the physical delivery of mail. Australia Post is less concerned with physical addresses that don’t receive mail, such as vacant land.
Australia has two national standards for address in operation: AS/NZS 4819:2011 and AS4590:2006. However, neither of these are mandated. When it comes to the differences between official principal address and the addresses that people use, there are some simple issues that impede the ability to accurately identify an authoritative address for a property or location.
Properties with multiple street frontages
The most common example involves the property located on a street corner. It is not unusual for such a property to have multiple addresses recorded for it, using numbering based on the different street frontages.
Alias and vanity addresses
This is where an individual address is also known by another name. An example of this is where a street address is provided against an alternative locality. This quite often occurs where an adjoining locality is seen as more prestigious than the suburb where the address is actually located. Another example is where locality boundaries do not coincide with major roads that are commonly perceived as marking out the locality. A new development may also adopt and market a name that is not officially recognised as a locality.
Incorrect names and types
Mismatches can occur due to incorrect or alternative spellings or where the wrong road type is used, for example, ‘avenue’ instead of ‘crescent’. One of the most common sources is where a major road or highway passes through a community and a local name is used along that particular stretch of road. For example, the Princes Highway is referred to as High Street when it passes through a particular hamlet.
Addressing within private developments
There has often been a lack of street addresses recorded within private developments such as gated communities, shopping centres, caravan parks and aged care facilities. As the local council is unlikely to have responsibility for the road network or provision of services within these developments, the address recorded reflects the local council’s main interest in the property, that is, as a single entity for rating purposes. A similar situation also arises with facilities such as military bases that are owned by the Commonwealth Government. Even though the military base or other facilities may include significant residential accommodation, these are not under the control of local government and are often not captured in state or territory government address files.
Complex sites and multi-unit development
Different uses of prefixes and suffixes for numbers, flats, level and building numbers also cause inconsistencies. For example, the address 1/17 George Street may be recorded by the jurisdiction but the mailbox may show the address 17A George Street. Ranged addresses (for example, 18-26 Railway Parade) are also a particular issue. These ranges may be allocated by a local council to, say, allow for the potential redevelopment of the property. However, most property owners tend to use the first number only as the address (for example, 18 Railway Parade). In the case of multi-unit developments, different buildings on the site may also use different numbers (for example, 18 Railway Parade, 22 Railway Parade, 26 Railway Parade) but other addresses for the property are recorded as 18-26 Railway Parade).
Historical or retired addresses
These are addresses that remain in use but may have been superseded for reasons including the re-subdivision of land; the introduction of rural addressing; the use of street numbers instead of lot numbers; renumbering of streets by local government; and the use of RMB numbers as part of an address. Disused mining towns are an example of where historical addresses may persist for whole communities long after they have fallen out of use.
The provision of addresses can also be a matter of timing. A development may be proposed and already have addresses allocated but work on the development has not actually commenced. Or it may take some time for the local government to capture addresses for a new development, although some parts may have started to be occupied with services being delivered to these properties. It is often difficult to distinguish between historical addresses and issues of timing.
Location of geocodes
Many large development sites contain multiple buildings and many addresses which often reference multiple street names. However, only one geocode may exist for the site, that is, a geocode captured at parcel level before the site was developed. Ideally the geocoding should reflect individual buildings on the site together with the allocation of the addresses to the appropriate building as required. This is also an issue for rural addresses where the geocode may not reflect the location of the actual residence on the property. There is also na increasing demand for geocodes to be applied to levels in multi-storey buildings.
Places without an address
We are also seeing an evolution in the understanding of what constitutes an address. Addresses originally developed around properties and were related to where a bill should be sent for the collection of rates or the provision of services by utilities. Land registries have progressively added into their systems, addresses for managing parcels of land. Now we are seeing addresses required to label not only properties and parcels of land but also assets such as community facilities, automatic teller machines, tree planting reserves and substations.
The governments of Australia established PSMA Australia so that the benefits of managing information by location (such as address) could be more easily accessed. This includes establishing national infrastructure that delivers a single point of truth and reduces inefficiencies.
PSMA Australia launched G-NAF, a geocoded national address file for Australia in 2004. G-NAF combines addresses provided by the land registries of the state and territory governments, the Australian Electoral Commission and Australia Post. The methodology used to build G-NAF also draws on the Gazetteer of Australia and other PSMA Australia-managed datasets for systematic validation. This includes the national transport dataset to provide data such as road centrelines and a component of the Administrative Boundaries dataset to provide information such as localities. This ensures high level of completeness in terms of coverage of all the addresses in Australia as well as relatively high levels of geocoding reliability.
In summary, G-NAF’s features include:
Since its launch, PSMA Australia has worked closely with the address contributors to continuously improve G-NAF. This has involved providing feedback and undertaking investigations to develop a better understanding of the complexity of issues associated with addressing in Australia. Substantial quality improvements have been made by the focused efforts of the contributing organisations in their source address datasets as well as the implementation of rules as part of PSMA Australia’s data processing.
Data contributors have been working to implement improvement programs to fix mismatches caused by alias names, incorrect spelling or incorrect road types. Governments are also in the process of strengthening legislation and other tools to compel developers to register official addresses.
PSMA Australia has actively developed rules to minimise duplicated addresses and has more than 200,000 in place to manage the update process. Issues arising from ranged addresses and from the use of levels have been a major focus. These are expected to be substantially reduced by August 2012. The inconsistent use of prefixes and suffixes is quite difficult to resolve due to the great variability involved. However, improvements through the use of rules will be implemented by November 2012.
Considerable programs are being undertaken by the state and territory governments to improve geocodes for new or recent developments as well as for rural and remote addresses. This includes a focus on complex addresses for multi-dwelling sites and private developments such as retirement villages.
The Australian Electoral Commission is also implementing a major project to spatially enable the electoral roll. Rollmap is a web mapping service that enables AEC personnel to use their local knowledge to assist in reducing the number of street level geocoded addresses by geocoding the household. This project will also provide feedback on potential aliases for addresses as well as on inconsistencies in addresses or geocoded locations.
Expanding address types
With the expansion of address types, PSMA Australia is using a working definition of an address as: A structured label for any location where one would expect to deliver/receive a good or a service.
PSMA Australia has introduced the Features of Interest dataset that contains many places not described by an official address. For example, urban parks, water storage facilities, national park camping grounds or boat ramps. The dataset incorporates comprehensive urban centre locality data, including indigenous locations.
Features of Interest is an independent dataset that can be integrated with associated datasets such as G-NAF, providing enhanced functionality. The dataset contains persistent identifiers and includes the following geometry data types: point (such as address), line (roads) and polygon (land parcel or boundary area).
The dataset is developed from feature data provided by government as well as selected organisations. The dataset includes multiple geometry types, and represents six main to level categories underpinned by more than 500 feature types.
Many organisations need to be able to search for new addresses as soon as they are created and lags in the provision of this information impedes the provision of services to new properties, such as utilities, telecommunications, constructions, insurance and marketing organisations. These delays particularly affect disaster and emergency management and also impact on other public services such as electoral rolls, vehicle registration, licenses and permits and the provision of health and welfares services.
With the quality of G-NAF improving significantly, PSMA Australia’s focus has turned to improving the access to up-to-date addresses. There is a significant amount of churn in addresses, with an average of more than 50,000 new addresses being added to G-NAF every three months.
The important commitment made by the governments of Australia to adopting a national coordinated approach to address management has created the right environment for this initiative. The National Address Management Framework (NAMF) is a practical, consistent, standards-based framework to guide the process for verifying and exchanging address information.
The goal of NAMF is ‘one address = one location’ and it guides the process for verifying and exchanging address information. NAMF also embodies critical concept of notification. This means that if a consumer of a NAMF compliant address validation service encounters an address that cannot be validated, it must be notified back to the data custodian. This provides user generated content that is both focused and verified by virtue of the framework’s operation.
G-NAF has been selected to support NAMF for location address operations and functions including address validation. To achieve this, PSMA Australia is working though the challenges associated with delivering this dataset more frequently to be better attuned to the needs of businesses and government.
This will be achieved with the release of G-NAF Live in late 2012. This is a web service developed by PMSA Australia and made available through the technology platform provided by PSMA Systems. Developed on standards-based service integration, PSMA Systems provides an easy, simple and cost-effective way to access PSMA data with very little development time, no capital investment and low operational costs based on transaction volume.
G-NAF Live will provide near real-time access to new addresses gazetted by state and territory governments. New addresses provided by contributors will be matched and corrected, then made available in a NAMF format with updates being provided on a daily basis. Most importantly, the notification process provides the ability for all users as well as the wider community to benefit from a wrong address being made corrected.
The implementation of NAMF and the launch of G-NAF Live provides real opportunities for the emergency management sector to realise cost and resource savings while enhancing operational efficiency and public safety.
The emergency management sector is in the business of understanding risk. If you consider that the natural and social environments we occupy contain a number of risks, disaster and emergency managers apply finances, resources and equipment to identify, mitigate and even treat these risks. Information management is one of the many important tools required to assist with not only the identification process but also response modelling and ordinary operations. Information that would allow the identification of the number and location of affected addresses and most effective travel route planning are examples of how any response could be optimised in terms of the resources required.
Even in situations where a risk has been modelled and even treated, the risk may still materialise into a disaster event but it would be expected that the impact would be lower. Now consider the alternate position where risks have not been identified, modelled and treated because of the limits on the different types of resources available. The limitations could apply because the scale of the risk is too difficult to understand, time required to model it isn’t available, the people skills are simply not present or ultimately the financial resources are just not accessible. It is inevitable that these untreated risks will materialise at some point, with significant impact for the environment and community. So the question for emergency managers is how can more risks go through the treatment process? The obvious answer is more effective use of the available resources, which in this case means (spatial) information management.
The limitations that exist in relation to available resources are a key factor in the distinction between treated and residual risks. The collective goal of the disaster and emergency services sector is to free up resources for the treatment of residual risks.
Diagram 1: Treatment of risks
Disaster and emergency management sector are facing increasing complexity around the compilation of the critical information for relevant operational systems. Isolating the task of data loading alone, it is not a simple activity and is time consuming. It is made even more complex when multiple sources are being loaded to produce the same information, such as in the case of address.
Looking across the sector, the duplication of effort required for all disaster and emergency service organisations to go through the process of managing address information is significant. Not only are there different systems in use across the various agencies, but this information isn’t being used in a standardised way. It is not unusual for an agency to have a compilation process that would take a number of weeks (perhaps as many as six to eight weeks) to produce their own products. Much of this time is spent simply loading, formatting and filtering data.
Real time address management through G-NAF Live would assist the sector by:
If real time address information is important to the sector, then a services approach to address management makes more sense. Even if incremental data feeds can acceptably be applied, there will always be a duplication problem as all users have to apply the incremental.
It is well known that the disaster and emergency management sector is fantastic at providing valuable feedback to address custodians when there are data issues. However, these notifications currently come in different ways and, in many cases, are based on personal relationships between the agency and government which increases the risk of inefficiency and error. There is also likely to be more duplication of effort when an address can’t be found. NAMF standardises the means for providing feedback as well as ensuring that the information is enacted.
There is also considerable opportunity to use post-incident analysis of location data to provide valuable additional information about how the community talks about their location. Through the NAMF notification process, this could be fed back into custodian addressing databases to help deal with some of the issues mentioned above.
G-NAF has undergone significant improvements since its launch in 2004. Many of these improvement programs, including those tackling the most circuitous issues, will culminate in 2012. The introduction of the Features of Interest dataset will also significantly broaden the range of places that are provided with an address.
The operation of NAMF with its critical focus on notification is a significant driver of improvements to addresses in Australia. The launch of G-NAF Live provides real-time access to new addresses that offers the disaster and emergency management sector the capacity to reduce duplication and improve risk management through freeing up of resources and funds. The industry’s participation in the notification process also has the potential to get official address right as well as enhance our understanding of the addresses that people use.
A significant challenge remains to engage with the largely overseas-based CAD vendor community to ensure that the increased data flow volumes are realised in improved operational efficiency.