Safety in Numbers
by Deirdre Kelly
photography by Sofie Kirk
A new analytics dashboard created by data scientists at York University’s Schulich School of Business will help pandemic decision-makers predict – with a high degree of accuracy – when to lift lockdowns.
Using publicly available data from the national health authority, the COVID-19 dynamics dashboard is an analytics decoder that can specifically pinpoint how fast the coronavirus is spreading and, conversely, how much the spread is decelerating as a result of expanded vaccination rollouts.
“Given how rapidly the COVID-19 situation is changing, this predictive data will be very helpful to policymakers, health-care administrators and public officials in determining when and where to ease lockdowns,” says Murat Kristal, director of Schulich’s Master of Management in Artificial Intelligence and Master of Business Analytics programs.
The performance metrics measure daily percentage change velocity and percentage change acceleration. The numbers provided by the dashboard are highly accurate, more than 90 per cent for a one-day reading.
The results are colour-coded.
Unsurprisingly, when the dashboard flashes red, it means that the virus is spreading; when it shows green, the speed of contagion has decreased, compared to the previous day’s update.
Currently, according to the dashboard, spread of the virus is slowing down.
“With the introduction of the vaccines, the acceleration measurement has switched from red to green,” Kristal says. “But in order to get out of lockdowns completely, we still need to have the speed turn green as well. I’m optimistic we will get there.”
Kristal developed the dashboard in collaboration with other data scientists at York, among them visual analytics and modelling instructor David Elsner (BA ’02; EMBA ’13) and Deloitte data scientist Hjalmar Turesson, a lecturer in Schulich’s Master of Management in Artificial Intelligence and Master of Business Analytics programs.
“The dashboard can inform people’s actions and help in the development of strategies that support the public good,” Kristal says. “It adapts to the data to show the impact of the removal of the restrictions.” ■