For years, we’ve been told that data holds the solution to most problems. Data is the new oil; whoever has the most of it and can extract the relevant information out of it wins. Companies and public sector organisations were encouraged to go deep into their archives and dig out all relevant data they could find. Troves of historical data was set be used for training and validation for machine learning algorithms, giving them the ability to predict the future with confidence. But then Covid-19 turned up.
Historical data is valuable because you can learn from patterns of the past to predict the future – unless that future is entangled in a global pandemic. “Everything is completely disrupted,” says Nadun Muthukumarana, public sector and transport Analytics lead partner at Deloitte, explaining that there’s still no clear indication of what the future holds. “The one thing that is for sure,” he says, “is that things will continue to keep changing.”
We don’t know when the situation will stabilise, or what the world will look like when it does, so the historical datasets we’ve come to heavily depend on have stopped being reliable. “A lot of people have been saying that the tools they were using before the pandemic can no longer be sufficient to predict in the new normal,” says Muthukumarana. This doesn’t mean we should ditch historic data entirely, but organisations need to find more contemporary datasets. “You almost don’t have a choice,” he says. “You need to use highly dynamic, recent data; we call it effervescent data, because it’s always fresh and new.”
The same rules still apply: the more relevant data, the better. Organisations only have a finite amount of effervescent data within their own systems, so to source more, they need to look across the public sector, within different departments that provide services to citizens for new datasets.
This is where the public sector has an advantage: the ability to collaborate. Unlike the world of big business, public sector organisations aren’t battling it out for clicks and cash, so a hugely broad range of organisations would be free to compile their effervescent data. “We are talking about health organisations sharing data with welfare, police and justice, transport – with the public’s consent and trust, of course,” says Muthukumarana. This would likely give them access to a plethora of insights to improve public services, creating a well-rounded picture of the changing needs of citizens.
When you pair this effervescent data with digital twins, you can use simulation techniques to provide all possible outcomes for requisite scenarios. The digital twin replicates whatever is going on in the real world, physical and non-physical, in digital form. Then you can use the effervescent data, along with the complex problem-solving abilities of AI, to simulate potential future scenarios. By combining the physical limitations of the real world with mathematics, these simulations can be incredibly accurate in their short-term predictions.
Deloitte’s Motion Simulator puts these concepts into action. Transport and supply chain operators have used it to improve the performance of their networks. By running daily timetables through simulations in their network’s digital twin, everything from bottlenecks and staff shortages to lack of capacity can be identified and fixed before the services even start running. These simulations also do an “asset criticality” calculation, which essentially determines which pieces of equipment you can’t afford to have fail, because if they do, the whole system will fall apart.
“It gives the operator a brand-new way of looking at their assets for the first time,” says Muthukumarana. “You couldn’t do that before we combined the operational data with the asset data in a unique way.”
Motion Simulator was developed before the pandemic set in, using historical data; but it was built to be flexible. “The technologies we use for Motion Simulator are based on gaming technologies,” says Muthukumarana, explaining that gaming tech is naturally dynamic because it needs to work for so many different player scenarios. The Motion Simulator infrastructure is just as adaptable: “We just needed to plug in the effervescent datasets in place of the historic data to predict accurate scenarios,” he says. And its capabilities aren’t limited to transport.
Digital twins can be created for anything, even whole towns or regions. Effervescent data about a specific area – how its construction and retail industries are doing, for example – can forecast a future downturn. “Social welfare organisations could potentially pre-empt a region going into poverty,” says Muthukumarana. “So rather than responding after it’s happened, they can put a support structure in place before the businesses close down. “This is what the future of public services could look like – proactive and personalised.”
The short-term nature of effervescent data means that these predictive simulations need to be run regularly, with constantly refreshed data. This will give organisations an accurate “what if” scenario that they can act on in the short term. For longer term projects, the Motion Simulator can be used for validation. If a capital infrastructure project has been called into question because of the changing socio-economic patterns post-pandemic, the simulator can run the options.
“When you do a long-term simulation, you’re not looking for an accurate prediction of the most likely scenario, you’re looking for the broadest range of scenarios,” says Muthukumarana. The Motion Simulator will run millions of different simulations through the digital twin to show the whole range of possible outcomes. “It will allow a strategic planner to see everything from the best possible scenarios to all of the worst possible options,” he says. “Based on that, they can do risk and contingency planning.”
This kind of simulation doesn’t just allow you to plan for the next worst-case-scenario the world has scheduled for us, it can also reveal what’s possible going beyond the capabilities of human imagination. It shows you all the options, even ones you may not have previously considered – or had discounted as too risky to try.
These technologies are already being used by some private companies, but Muthukumarana expects public sector organisations to increasingly adopt this technology to make accurate predictions based on the rich data they already have. “If you look back, there are all sorts of examples of the public sector coming up with real digital innovation,” he says. “The modern-day software industry was born out of Nasa trying to get a man on the Moon.” Navigating a pandemic, and making it out the other side, will unearth innovations that will likely be just as influential.
–For more information, visit deloitte.co.uk/publicsectorsimulation