Brother's Art -

How Transport for NSW is tapping machine learning

Transport agency in New South Wales taps machine learning to restore public confidence in the state’s transportation network amid the pandemic

At the peak of the Covid-19 pandemic in 2020, Australian transport agency Transport for New South Wales (NSW) had to restore public confidence in the state’s transportation network and curb the spread of the disease.

One of the ways it did that was to analyse the travel history recorded by Opal transit cards – with an individual’s permission – and inform the commuter if the regular buses and train services that they had been taking were Covid-safe.

Chris Bennetts, executive director for digital product delivery at Transport for NSW, said those insights were derived using a machine learning model that predicts how full a bus or train carriage was going to be at a given time.

Based on the predictions, commuters would be advised if they could continue using their regular services or switch to a different service or mode of transport. “That was interesting for us because it was our first foray into personalisation to offer more choices for customers,” said Bennetts.

While the initiative was spurred by the onset of the pandemic, Transport for NSW had already laid the foundation for its work in machine learning around three years ago when it started using Amazon SageMaker to make predictions around commuter patronage.

“We were beginning to collect data in real time around commuters on the Opal network and using that data to make predictions about the impact of weather, special events and public holidays on the network, and applying machine learning to drive transport planning outcomes,” Bennetts said.

Transport for NSW also worked with a startup firm that has been tracking real-time movement of public transport vehicles on Google Maps, along with the number of commuters on each bus or train service.

Bennetts said the agency has since partnered with the firm to layer on public health information, such as the number of Covid-19 cases, to build a “layered picture” that central government teams can use to support decision-making.

Having open access to data is key to any analytics and machine learning initiative. Bennetts said the NSW government took a progressive stand by opening access to its data in 2016 through its open data platform that houses more than 15,000 datasets to date.

“Since the time that we launched it, we’ve hit 9.4 billion API [application programming interface] calls,” he said. “We’ve got 45,000 developers using it, with around 5,000 applications using the data.”

In the case of Transport for NSW, however, Bennetts stressed that the agency has been clear with the developer community that the data offers insights about what’s happening across the transportation network – and not personal data about commuters.

 “And we tend to wrap innovation challenges around that data as well to help promote the ecosystem thinking about what’s possible with these datasets,” he added.

Building an ecosystem around the use of open access data to build machine learning models, said Bennetts, was a better approach than doing it internally within the agency which did not have in-house capabilities to do so.

“What we’ve found over the past couple of years is that the startup community is really getting smarter about how to apply this stuff,” Bennetts said. “And so, we partner hard with the startup community and see what’s possible.”

As one of the world’s leading transport planners whose Future transport technology roadmap 2016 became a seminal blueprint for its peers in other countries, Transport for NSW is now leveraging Australia’s position as one of the forerunners in quantum computing.

Last month, it announced a quantum computing research project with Australian company Q-CTRL, a spin-off from the University of Sydney’s quantum science group, to study how quantum computing can create and manage a more resilient transport network.

Future applications of the technology could include mapping all transport modes, including autonomous vehicles, and crowd movements simultaneously in real-time, enabling transport schedules to be automatically updated to prevent service disruptions.

“If you have a known set of inputs and allow for a set of output possibilities, the quantum algorithm you can apply to look at all the exponential possibilities is well suited to solve problems around traffic management and dynamic scheduling of transport services,” Bennetts said.

However, the number of qubits – basic units of information used by quantum computers – available today is nowhere near what is required, he added, noting that the agency is looking to grow and scale the project, paving the way for it to become a world leader in applying quantum computing when the technology matures.

Read more about machine learning in Australia

Read more on Artificial intelligence, automation and robotics

Data Center
Data Management