Big data from RPI says local COVID-19 cases could peak at end of May if social distancing remains

PHOTOGRAPHER: youtube

Rensselaer Polytechnic Institute

A big-data expert at Rensselaer Polytechnic Institute has developed a computer model that predicts coronavirus cases in the Capital Region will peak on May 29 – if social distancing measures are still largely in place.

Professor Malik Magdon-Ismail said he is confident in his prediction because it combines hundreds of separate models for how the virus can play out, rather than relying on just one theory.

“Many models are consistent with the data, and each of those models gives you a different prediction. So, it’s important to know what’s the range of things that can happen,” said Magdon-Ismail, a professor in RPI’s computer science department.

GAZETTE COVID-19 COVERAGE

The Daily Gazette is committed to keeping our community safe and informed and is offering our COVID-19 coverage to you free.
Our subscribers help us bring this information to you. Please consider a subscription at DailyGazette.com/Subscribe to help support these efforts.
Thank You

Since the 1990s, Magdon-Ismail has been studying the remarkable power of computers: often they can understand and spot trends better and faster than humans. While watching pundits on ABC discuss the spread of the virus in March, he realized he could use his number-crunching experience for good.

“That got me thinking that it’s good to have all this public health opinion, but what does the data say?” he said. “I’m not on the front lines, so what I can do to help is to help understand what’s going on from the data.”

Pulling new COVID-19 positive test results from the New York State Department of Health in Saratoga, Schenectady, Albany, and Rensselaer counties, he lets a powerful computer synthesize hundreds of models that can fit the existing data. His program estimates the Capital Region will peak on May 28 at 29,000 infected.

This total would include asymptomatic carriers, or apparently healthy people who spread the virus. It also assumes 75 percent of the Capital Region is still staying at home, per orders of Gov. Andrew Cuomo.

But his program could also help policy makers decide how and when to open their local areas. For example, if only half of the Capital Region stays at home, the peak would be predicted at 58,000 cases on June 8.

“It’s important to make it local, because what’s going on in Albany is very different to what’s going on in NYC,” said Magdon-Ismail.

So far, no one from New York State, the city of Albany, or other localities are working with Magdon-Ismail on public health decisions. But he’s hopeful that his research can help; his machine learning program could analyze data from any small city or region in the world, he said.

“One of the things that officials could say is, ‘What would happen if we relaxed social distancing levels by 50 percent?’” Magdon-Ismail said.

Cuomo announced Thursday that New York’s ‘on PAUSE’ regulations will be extended until May 15. He also said his office was planning how and when to reopen the state economy.