Locate22 Keynote Address: Building geospatial data capabilities to inform Australia’s important decisions

Dr David Gruen AO
Australian Statistician
Wednesday 25 May 2022
Locate22 Keynote Address: Building geospatial data capabilities to inform Australia’s important decisions

Thank you to Dr James Johnston and the Locate22 organising committee for inviting me to speak at this preeminent gathering of the geospatial community in Australia. I would like to begin by acknowledging the Ngunnawal people the Traditional Custodians of the land we are meeting on today and pay my respects to their Elders both past, present and emerging.

As Australian Statistician and Head of the APS Data Profession, I’m here today to talk to you about lifting data capability across the APS and making better use of data – and in line with the focus of this conference I will give particular emphasis to geospatial capabilities. I’ll give you an overview of our role in the APS Data Profession; provide some examples of innovative APS data projects; outline ABS efforts to lift our location information offerings and geospatial technology capabilities; and some of our new partnership with Geoscience Australia and other APS agencies that will provide new location insights.

I thought I would begin by providing you with a little of my background. I started my career as a research scientist. Since then, I’ve undertaken many different roles spanning research, analysis, and policy development, including in the Reserve Bank of Australia, the Australian Treasury, the Department of the Prime Minister and Cabinet, and now as the agency head of the Australian Bureau of Statistics.

All these roles involved data in some way – using it myself, managing staff who use it, or more recently leading projects aimed at making better use of both public and private sector data. This included developing the Government response to the Productivity Commission inquiry into Data Availability and Use, which culminated in the Data Availability and Transparency Act recently passed by the parliament, and the Data Integration Partnership for Australia, a $130 million initiative to build data infrastructure and capability across the APS. More recently these initiatives are feeding into the development of the Australian Data Strategy.

In the leadup to this conference, I was interviewed for a podcast to promote the conference. I was asked about my first exposure to geospatial data. While I have undoubtedly encountered geospatial information many times in my career, a memorable example was developing the Drought Map in response to the millennial drought. The Drought Map provided a visual representation of the critical geospatial aspects of the drought, as well as the relevant community and government services – a visual representation that was both compelling and easy to understand.

More generally, location data and insights provide critical inputs for governments, business and the community to improve decision making at the local level – for example to support first responders to an emergency event or in preparation for Australia high-risk weather season.  These data also inform the best places to build critical services like hospitals, breast screening centres and health services for Aboriginal and Torres Strait Islander peoples.  They help inform investment in critical infrastructure for economic development or climate risk mitigation, regional economic development strategies, and the delivery of employment and education services and programs. Decisions informed by high-quality location data have a direct impact on community welfare.

Since the early 2000s, there has been rapid growth and diversification of the data ecosystem with the digital revolution creating a data deluge in its wake. People use digital platforms every day, generating huge volumes of data, much of which has a location reference associated with it. Digital technology has changed the way we work, shop, socialise, travel, and access services such as health care. Maps and location tools are part of the digital applications we now take for granted.

The increase in the types and amount of data from both the public and private sectors has led to rising demands for information from users. Being able to use data for policy development and service delivery is now an expected part of the efficient operation of the public service. We have responded with collaboration and partnerships across government and with the broader data community to build capability, tools, and processes to securely use, share, and understand data across the APS and with other governments. Through these collaborations we are also gaining access to new sources of information.

In order to make them most of this rapid growth in data availability, we need to build workforce data skills and capability across the APS. Working on data used to be predominantly the domain of statisticians, economists, geospatial analysts and meteorologists. These days, and increasingly, most roles in the public service (and indeed in many spheres of work) require some level of data literacy. The APS Data Profession is playing a major role uplifting workforce capability across the public service.

Before talking about the Data Profession in detail, I want to outline how the Profession came about.

In 2018, the Australian Government commissioned an independent review, led by David Thodey, to assess whether the APS was fit-for-purpose for the coming decades. The Terms of Reference for the review noted the APS needed to be apolitical and professional, agile, innovative, and efficient — driving both policy and implementation through coherent, collaborative, and whole-of-government approaches. The APS must have the capability to meet core responsibilities and deliver functions, and to understand and deploy technology and data to drive improvement.

The review concluded the APS should make better use of data and analytics to generate deeper insights, provide better advice to inform government decisions, and enable more effective service delivery and regulation to improve social and economic outcomes.

In response to the review recommendations, the Australian Public Service Commission (APSC) established three Professional streams in critical disciplines and functional areas. The Human Resources and Digital Professions were the first two launched.

The third was the Data Profession, launched in September 2020. I was appointed Head of Profession, with the ABS as lead agency. At the centre of the work program is the Data Profession Strategy, which addresses the need to strengthen data capability overall, as well as build niche expertise for specialist data users – with geospatial being a prime example of a specialist skill.  The strategy aims to lift the data capability of the APS workforce by defining data capabilities, increasing diversity and mobility of people in data roles, and creating career pathways and development opportunities. The Program aims to ensure the APS can attract, develop, and retain people with the data capabilities required to harness the unprecedented growth in the availability and value of data.

The APSC’s 2021 State of the Service Report identified data as the second most common skill shortage identified by public service agencies, and the fourth most common identified by employees. The Data Profession seeks to narrow, if not close, these capability gaps.  My role is to guide the Profession’s progress against the initiatives, priorities and success factors outlined in our strategy.

So far, the Data Profession has made good progress uplifting data capability across the public service in a relatively short time. Let me tell you about some of the initiatives. We established a Data Professional Network, to guide and share professional data standards and best practice, which now has 1,700 members.

On recruitment, the ABS leads a streamlined data graduate recruitment round on behalf of the APS, recruiting Data Analysts, Data Scientists and Statistical Methodologists. This year we are recruiting for 39 APS agencies and hope to fill about 300 positions. In a sign of the increasing importance of the skills represented at this conference, this year we included a dedicated stream for geospatial specialist graduates.  The applications have closed this year but for anyone who missed out they will reopen in March 2023 for the subsequent year’s intake.

This year has seen strong demand for graduates among Australian employers, which has resulted in a fall in data graduate applications from a high of 2,400 last year to nearly 1,700 this year.

Late last year, we collaborated with Geoscience Australia, the National Recovery and Resilience Agency, and the Department of Infrastructure, Transport, Regional Development and Communications to run a recruitment round specifically for geospatial specialists. This round attracted a strong field of 140 applicants from across industry and government, enabling us to fill over 20 geospatial specialist roles across the APS from operational analysts (APS5s) to Senior Managers (EL2s) with the majority coming from outside the APS.  Given the success of this recruitment round, we plan to run a similar process toward the end of 2022.

We’ve also supported a mobility program of immersive learning experiences, which develops specialist data capability for APS employees through job swaps or secondments. So far, we have facilitated 13 short-term arrangements, where APS staff with data expertise work temporarily in another agency or department.  Geospatial specialists have had a prominent place in these placements, with moves between ABS, Geoscience Australia, the Bureau of Meteorology, the Department of Agriculture Water and Environment, and the newly formed Australian Climate Service. When these staff return to their home agency, they have a better understanding of how data is used in another data-rich agency and better links to another data team.

One of the most important pieces of work the profession has undertaken is the development of an APS Data Capability Framework, which defines the data skills, knowledge and behaviours required to perform the myriad of data roles in the APS. What is valuable about the framework is that it elucidates and catalogues the wide range of data roles available across the APS.

When talking about data skills, many people think of data analysis, and of course data analysis is important. But the framework catalogues the wide variety of data capability areas including data management, data collection, data governance, data classification, and data communication and visualisation. The role of geospatial skills and knowledge is referenced in the framework and more detail about these roles is emerging as the framework is being used. The Beta version of the Framework is available on the APSC website.

I want to turn now to some innovative data projects being undertaken across the APS, because while we recognise the need to lift data capability, the APS already has highly skilled and experienced data professionals doing important things with geospatial data.

My examples are about projects that involve the ABS, often in close collaboration with others. I choose these examples because I know them best, not because the ABS has a monopoly on innovative data projects with a geospatial dimension across the APS – we clearly do not!

The first example involves our use of Australian Tax Office Single Touch Payroll (STP) data to provide insights about the changing level of jobs and wages during the COVID lockdowns across different parts of Australia.

The STP dataset contains de-identified data on 10 million employees – which is close to a census of payroll jobs.  This allows us to produce detailed geospatial analysis that is not possible using the 50,000 or so individuals from whom we collect data in the monthly Labour Force Survey – this coverage and detail is one of the big benefits of administrative data sources.[1]

ABS received its first delivery of STP data from ATO in early April 2020 in the first stages of the COVID pandemic. In under three weeks we produced data for the first release of a new publication, ‘Weekly Payroll Jobs and Wages in Australia’, which contained aggregate jobs and wages data for the Commonwealth as well as for States and Territories.

Soon after, in mid-May 2020, we released data for our Statistical Area 4 (SA4) geography, which is a broad regional level geography. This data included payroll jobs indexed to the number of jobs in the week of 14 March 2020 (when Australia reached 100 cases of COVID-19). We also produced an interactive map to provide a visual representation of changes in these payroll jobs across Australia. This was the first opportunity government, business and the community had to see the localised impacts of COVID, based on where people lived.

We then went to a finer level of geographic detail with a September 2020 release of payroll job indexes at the Statistical Area 3 level.  This was shortly followed by an interactive map showing changes in payroll jobs at the SA3 level that gave policymakers, and the community more broadly, a more geographically detailed picture of the impact of COVID on jobs.

ABS continues to release Payroll Jobs and Wages data from this valuable source of data from the ATO. It is helping analysts track the jobs recovery and the emerging pressures in the labour market.  Recently we released new SA3 level maps that showed how jobs in New South Wales and Queensland had been affected by the recent severe floods, giving near real-time insights into these localised impacts.

This use of this Payroll data links to my next example where we are an active partner in the Australian Climate Service, or ACS for short. The ACS was established last year as a virtual partnership, bringing together expertise and data from the ABS, the Bureau of Meteorology, CSIRO, and Geoscience Australia. The role of the ACS is to help the government and the community better understand the threats posed by natural disasters, including those that have been intensified by climate change, and limit their impacts now and in the future. The ACS connects the Commonwealth’s extensive climate and natural hazard information into a single authoritative and detailed source of truth.

The ACS delivers data, expertise, and advice to two customers: the newly formed National Recovery and Resilience Agency (NRRA) and Emergency Management Australia (EMA). The most recent floods in Queensland and NSW saw ACS provide rapid response to data requests to help these agencies understand developments on the ground.

As part of the ACS effort, ABS provided data to NRRA within 24 hours to help them estimate the population adversely affected by the floods. The request came in on a Friday evening, to feed into an urgent briefing that weekend. We mobilised quickly, established data sharing agreements, prepared the data, and set up a secure transfer mechanism to get the data to NRRA by early Saturday afternoon.

The data was based on the ABS Population Grid, which combines the latest ABS regional population estimates with an enhanced version of the 2020 Geocoded National Address File (or G-NAF) from Geoscape Australia – who also provided quick agreement to share the enhanced data.  ABS provided estimates of population at a Statistical Area 1 (SA1) level modelled down to the address level. NRRA then used these point estimates of population to dynamically estimate the number of people affected, by matching the points to flood extents as they moved and changed.[2]

My third example is work we’ve been doing using existing public and private sector administrative data sources to generate new monthly indicators of economic activity.[3]

The first output from this project was a new indicator of business turnover, based on Business Activity Statements submitted to the ATO. The first release was in October 2021, and monthly since then.

We began the release of a second monthly indicator in February – this time providing a measure of household consumption, based on aggregated, de-identified transactions data from Australia’s major banks.

Household consumption accounts for about half of GDP, so there is considerable value in having an accurate measure of it. The existing monthly measure of household consumption is based on the Retail Trade Survey, which covers about 30% of household consumption.

The new measure, based on banks’ transactions data, covers 68% of household consumption, so that is a substantial step up. Many items of household consumption are captured in the new transactions-based measure but are missing from the retail trade survey. To give a few examples: purchase of petrol, car servicing and maintenance, train and bus tickets, Uber rides, airline tickets, hotels, theme parks, haircuts, dentists and allied health costs. You get the idea!

Next year we will also be releasing a new monthly earnings indicator using Single Touch Payroll data from the ATO.

The new indicators are currently being produced at broad levels, such as by industry division and at national or state/territory levels. In future, we plan to explore providing these data at finer levels of geography than the current State-based estimates. In doing so, we need to ensure individuals’ privacy and respect the commercial interests of the data owners.

My final example is a project for which ABS received funding in the most recent Federal Budget to produce enhanced regional labour market statistics. With this funding, ABS will produce more reliable monthly measures of regional labour markets – that is SA4 measures of employment, unemployment, underemployment and labour force participation across Australia. From 2024-25, ABS will also deliver new quarterly and annual labour market statistics for sub-regional (SA3) and suburb/rural locality (SA2) geographic areas, and Local Government Areas. The labour market statistics will extend coverage to target population groups, such as employment for women.

ABS will use new data science and statistical modelling methods to combine existing survey and new administrative data sources to produce this new labour market data at finer levels of geography than previously possible. We will also enhance this data by integrating it with geospatial systems, maps and dashboards, including the GA’s new Digital Atlas and DITRDC’s Regional Data Hub.

These are just a few examples of what the APS is currently doing with data to integrate different data sources, including geospatial data, to meet critical information needs. It is worth reflecting that not so long ago we couldn’t do these things because there was much less sharing of data between agencies; the supply of data was more limited because digital platforms had not reached a level of maturity that enabled the generation and integration of large amounts of data; and we didn’t have ready access to the required technology, such as cloud storage.

We’ve come a long way but undoubtedly, we can do much more provided we continue to lift data capability across the APS and in the wider community. The passage of the Data Availability and Transparency Act in March 2022 is also an important part of this journey and will be a key enabler supporting our future efforts.

Also of interest to this audience is the work we are doing at the ABS to uplift our own location information and geospatial technology capabilities. This has been in response to increased interest by our key customers in using location data and local area insights to identify and respond to economic, social and environmental challenges. The COVID-19 pandemic and recent natural disasters have further raised our customers’ interest in location data and analysis by demonstrating their invaluable role in responding to challenges with a significant geographical dimension.

The ABS has always provided much local and regional socio-economic data for use in location analysis. Our Population and Agricultural Censuses have been rich sources of data on local communities and for small population groups. New data from these censuses will be released later in 2022 – with the first tranche of Census data to be released on June 28.  I suspect many of you are eagerly awaiting the new small area census data that will become available then.

The ABS is actively working to go beyond these traditional sources of socio-economic small area data. As a first step we created a new Location Insights Branch in June 2021. It is a virtual team that now includes over 70 staff located in Canberra, Adelaide, Melbourne, Brisbane and Hobart. This branch covers our new work on the Australian Climate Service and Digital Atlas – which James Johnson will talk more about in the next session.  The new Location Insights Branch consolidates ABS’s existing location and geospatial programs – our statistical geography, geospatial analysis and technology area, our statistical address register, and our team that compiles a wealth of small area socio-economic statistics.

This branch will play a key role in ABS’s ambition to use the new data sources I mentioned earlier to deliver additional, timely location-based insights for our economy, people and environment.  One example where we are already exploiting geospatial data is our use of Earth Observation (satellite) data products form Geoscience Australia and ABARES. We have used these data to compile and release the first detailed set of National Land Accounts, which will inform environmental decision making across Australia.

This branch will also feed these new and existing data sources into web-based data services for the Australian Climate Service and other users. A focus here is to ensure the data gets into the hands of the Emergency Management Australia and the National Recovery and Resilience Agency to enable better planning and preparedness for natural hazards and disaster response and recovery. In addition, we will be working with Geoscience Australia on the Digital Atlas program to increase the amount of local area socio-economic and geospatial data that is readily available and accessible for use in location analysis and decision making.

I’ve talked about developments in the APS Data Profession and given you some examples of innovative data projects that are providing new sources of geospatial data for our range of customers and will provide new data for broader geospatial analysis and research. I have highlighted ABS’s commitment to developing geospatial data and capabilities, so that we can supply you with the right data in the right geospatial formats. The APS is building our data workforce and technological capabilities to support these objectives.

Ultimately the aim is to inform decisions at a local level by bringing socio-economic data together with geospatial data, tools and knowledge. We welcome new people, ideas, partnerships and collaboration to support the goals of protecting and enhancing peoples’ lives, helping our economy to adapt and prosper, and caring for our environment assets.

ABS is at an exciting stage in our evolution. We have new data coming from our traditional Population and Agricultural Censuses.  And we will soon be supplying data in new ways, including via geospatial enabled web services and through Digital Atlas. Our use of new data sources, including private sector sources, means richer, more detailed data, for smaller geographic areas. This will provide you with opportunities to ask new questions and to get answers to questions you could not previously answer.

In time, you will be able to develop new insights for your favourite geographic region – whether it’s a suburb, river basin or electorate, or for a bespoke region like a 100km zone from the coastline. You will be able to use our geospatial enabled socio-economic information to inform the decisions of your clients, your organisations and your communities.

Thank you.

[1] With data on 10 million employees, disaggregated analysis can also be presented by age and gender and by industry as well as geospatially.

[2] The data provided detailed local information without compromising the privacy of individuals or households. G-NAF is a public dataset of addresses and by modelling population using SA1 level data the ABS protects individuals’ privacy. There are between 200 and 800 people in each SA1, and about 62,000 SA1s across Australia.

[3] A significant benefit of using existing data, collected for other purposes, to generate these new indicators, is that there is no need to put a new survey in the field, which would unavoidably place a burden on respondents to the survey.