Getting started with GNAF - Hugh Saalmans

Interview with Hugh Saalmans, Location Science Manager at IAG. Hugh or @minus34 is an open data advocate and an expert user of GNAF and Administrative Boundaries. In this interview, he provides some background and advice for getting started with these datasets.

Tell me about your areas of expertise, your past and present roles, and the work you do.

I’ve been working with geo-information & tech for over 20 years. I’ve had roles in pre-sales, consulting and product management with MapInfo (Pitney Bowes), RP Data (CoreLogic), Intergraph and MapData Services. I switched over to the customer side over 5 years ago and have been embedding geo-information across IAG, Australia & New Zealand’s largest general insurer, ever since.

I’ve worked for a number of years with address validation & geocoding, route optimisation, mobile workforce management, web & mobile mapping, system architecture; using a host of open source and licensed technologies.

Right now, I lead a team at IAG that’s putting all our corporate data on the map (keyed by G-NAF PID or persistent identifier) and making it available as APIs for analytics and visualisation across IAG.

How did you become familiar with G-NAF? How long have you been using G-NAF?

I’ve been using G-NAF heavily for 10 years; mainly using it for address validation & geocoding as well as spatial analysis; usually on a national scale.

I’ve also used the PSMA Administrative Boundaries for over 5 years for data analysis, aggregation and visualisation.

What does G-NAF help your organisation to do and achieve?

The fundamental purpose of G-NAF is to enable the validation and geocoding of physical addresses.  This underpins any economic activity that could benefit from a knowledge of where users, customers or citizens are, along with their verified address.

In the insurance industry we use it to locate customers to the property level across Australia; enabling a number of core activities. The two most notable being:

  • Pricing several million customers based on their individual risk, rather than charging one rate for each suburb or postcode, and
  • Understanding the impact of a disaster on our customers; so we can respond as quickly as possible in their time of need.

Why is it important for Australia to have a geocoded national address file?

A national reference set of geocoded addresses is fundamental to the delivery of products & services to consumers & citizens; as well as the analysis of trends at a granular level.

There are many organisations that benefit from validating and locating an address to do business, and to understand their customers better. Some good examples include:

  • Banks assessing a home loan application
  • Insurers verifying  a property and pricing its risk appropriately, and
  • Utilities delivering water, gas, electricity or Internet services.

Why is it good news that G-NAF is now being released as open data?

Making G-NAF and the Administrative Boundaries open is a great thing as they’re important datasets that can be used to drive innovation and economic activity in the digital age.

Also, it enables organisations to create new products, share their data and collaborate without any limitations on their use of G-NAF.  It also allows them to create, share and collaborate using a ‘standard’ Australian persistent identifier (the G-NAF PID).  In the past, this wouldn’t have been possible without all parties being licensed for G-NAF.

How might open access to G-NAF support Australia’s start-up, business, research and community sectors?

Removing licensing costs of G-NAF and limitations on derived products opens the data up to a number of great things:

  • Start-ups will be able to use G-NAF as the backbone of new, low cost, low risk business ventures
  • Consultants who previously couldn’t justify the cost of G-NAF for ‘one-off’ client jobs will now be able to do the work with minimal risk
  • Researchers will be able to conduct granular, national research at a lower cost using authoritative data
  • Charities and NGOs will be able to use data that was previously outside their budget for research and service delivery
  • Small to medium businesses will gain access to datasets that have a lot of business benefit, but might not have been cost effective previously
  • Corporates will be able to use the data more broadly across their organisations, and not limited to one area or registered business
  • Developers will get a set of rich, universal datasets that can be used as the foundation of their creative new concepts and apps at hackfests (such as GovHack) and other proving grounds.

The Australian Government’s Open Data 500 report found that geospatial data was the most popular form of government data being used. Why do you think this is the case?

Geospatial analysis and visualisation allows an organisation to discover trends that can’t be uncovered using numeric and textual analysis. A lot of organisations know this and use geospatial data to add the “Where” factor to improve their business.  As some of the biggest contributors to government open data in Australia are the Australian Bureau of Statistics, the Bureau of Meteorology, Geoscience Australia and a number of local governments – organisations whose work is heavily location based: there’s an abundance of open geospatial data available to support ‘geo-enabled’ organisations.

Do you have any advice for someone approaching G-NAF for the first time? 

Look for open source tools or pre-built databases on sites such as GitHub that will allow you load simplified versions of G-NAF without too much hassle. Otherwise you’ve got a bit of work to get started (31 tables to create and up to 144 pipe separated files to load).

If you want to work closely with the data, have a good read of the product documentation on the data.gov.au and PSMA website.  You’ll need to invest time to load the data yourself if you want to understand the raw data better.

Don’t get bogged down investigating every addressing quirk you find in your quality assurance.  There are lots of them in Australia and most of the time that’s just the way it is, as opposed to there being an error in G-NAF.  Australia’s addressing systems are good, but not perfect.

If you’re doing address level analysis or visualisation – make sure you pay special attention to the addresses with street and locality level geocodes.  They may give you unwanted results.  The good news is these make up <5% of all addresses.

Do you think that location data/GIS can make the world a better place?

You only have to look at the Open Street Map efforts in Nepal and Haiti in recent years, or the work the local MAPS Team or the Tomnod crowd do when disaster strikes; to see the positive impact geo-information has.

In the insurance space, when you can send a drone into a no-go area after a bushfire to assess properties from the air, you’re helping people who may have lost everything by giving them some certainty about their future.  The benefit of being able to use location data is very tangible.

Hugh's quick guide on G-NAF is available here or you can refer to PSMA's Getting Started Guides for G-NAF and Administrative Boundaries.

Products: 
G-NAF
Scroll to top