A digital government twin can model effects of change over time. It could teach us what structures worked and what didn’t, optimising machinery of government

By Allan Barger

Thursday July 18, 2019

Picture: Getty Images

Allan Barger is a digital government specialist who worked on data.gov.au and most recently in the New South Wales government. He shows how even a simple ‘digital twin’ — a virtual replica of a system that exists in the real world — can help to better understand the shape of government and start a conversation about how we can use technology to shape a better bureaucracy.

In many cases, even when a government stays the same, it changes. It does so for many reasons. Sometimes, ministers are promoted and given more responsibility; sometimes, it changes because it determines new priorities.

Whatever the reasons, such changes can have a colossal effect on the public service. Government agencies might find themselves moved, split, or even abolished. Their staff, tenancies and technologies can end up scattered across many receiving agencies. Costs will range from rebranding to rehoming and redundancies.

Insights & analysis that matter to you

Subscribe for only $5 a week


There has been some interest in the efficacy and costs of these changes to the machinery of government. A Victorian government inquiry found that: By the end of January 2016, the direct MoG [machinery of government] costs following the 2014 election had already exceeded $5 million.” Likewise, the Australian National Audit Office published a report on the costs and benefits of these changes based on information gathered about the machinery of government changes made in 2013 and 2014. 

I, too, am interested in thinking about the machinery of government. I have been striving to understand and find a way to holistically think about the shape of government. Recently, I have been really interested in the concept of digital twins.

A digital twin is a clone of a physical entity but in a digital space. Living digital simulations are used extensively in industrial engineering, and more recently has become popular for modelling urban sustainability.

The benefit of foresight rather than the misery of hindsight

A digital twin is not just a realistic three-dimensional representation of a system — a digital twin is a functional digitised model that takes into account how the system works and lets you observe the system in real-time or, alternatively, observe the effects of a hypothetical change. An example of a digital twin could be an office building reproduced in a digital setting that includes all the fixtures and fittings, and then uses sensors to recreate a day in its life. It lets you see how people come and go, what facilities they use, and what needs maintenance.

Collect enough data, and you can start to see how the building operates as a system. Where do all the meeting-room chairs go? Which lunchroom is the most popular? When do most people arrive, and when do they leave? These insights can help you understand and ultimately shape a better building because you can take a systematic view. Given enough ingenuity, you can even model the likely outcomes of change. What if we didn’t open until 9:30am? When is the optimum time to turn on the heating? Should we buy one of those umbrella-protector machines?

Digital twins aren’t just constrained to an office building. They don’t even need to be present in physical space. You can twin a pharmaceutical, the space shuttle, a city, or even a whole state. The New South Wales government, in partnership with the CSIRO’s Data61 digital research network, has been building a digital twin of itself that includes information on buildings, land, transport, and administrative boundaries. It’s a very good and worthwhile resource that will, without a doubt, prove invaluable for planning future infrastructure and livability-enhancing projects.

Leveraging digital twinning to plan the best machinery of government

We can take away something important from the concept of digital twinning and apply it to how we think about and represent the machinery of government: we can adopt a similar approach to what people have used to represent outbound functions, boundaries, assets, and materials, and apply them to agencies, ministries, legislation, policies, and people that make up a government. 

I’ve even had a go at doing it myself. I used the most recent Administrative Arrangement Orders, the Ministry List, and a copy of the very handy Finance FMA Flipchart to start mapping the landscape as it exists today, taking into account the data I could glean from the aforementioned documents.

I made the following ‘twins’ using a combination of neo4j and Google Docs. You can grab a copy of my code on GitHub.

This is what our federal governance landscape looks like today: 

A graphical representation of the relationships between ministers (light blue), ministries (red), Cabinet (sandy), parliamentary roles (green), departments (orange or dark blue), agencies (light pink), matters (yellow), and legislation (dark pink).

I can ask the program to do interesting things. Like showing me how the ministers and their ministries fit together:

A graphical representation of the relationships between the Cabinet (sandy), ministers (light blue), ministries (red), and departments (orange).

Or, to tell me all the departments that deal with a matter that includes the word ‘health’:


A graphical representation of the departments (dark blue) that deal with a matter that includes the word ‘health’ (yellow).

The (model) future is there for the taking

What could we do with a more comprehensive digital government twin? Could it help us understand how the existing system, composed of people, policy, legislation, and technology, works together to support the outcomes that a government wants? Could it help us shape the bureaucracy to better meet expectations of government and people into the future?

I think that by using a comprehensive digital government twin, we could find gaps, identify opportunities to improve and find clear areas of duplication in the system. More importantly, it will provide us with an understanding and ability to model how we can shape the system to best serve Australia and Australians into the long-term future. 

A digital twin could help us model their effects of change over time. We could learn about what structures have worked well and which didn’t. We could overlay information about their costs and make future machinery of government changes based on documented experience.

I built this simple ‘twin’ to better understand the shape of government and now I’m sharing it to start a conversation about how we can use technology to help shape a better bureaucracy in future.

About the author
Inline Feedbacks
View all comments

The essential resource for effective
public sector professionals