Thomas Kirchner and Debra Laefer
Thomas Kirchner (GPH) and Debra Laefer (TANDON)
Deploying Machine Learning to Understand the Virus’s Urban Spread
By Sarah Hagi
When the COVID-19 pandemic hit, the public was blindsided. With uncertainty about how the virus transmitted, how to stay safe, and exactly how our lives would change for the months to come, much remained a mystery. Working under an NSF Rapid Response Research grant, Thomas Kirchner of the School of Global Public Health and Debra Laefer of the Tandon School of Engineering were together able to gain some insight into how the COVID-19 virus spreads.
Using 3D data to capture human movement, the team investigated exposure and transmission patterns by mapping conditions around medical facilities and subways. Their research hopes to provide the groundwork to build machine-learning models to analyze how the virus spreads in urban areas.
Because COVID-19 hit New York City so hard, the duo had to act fast. Working under such pressure was incredibly intense. “I cannot begin to tell you how difficult this was,” Laefer says. “It was the craziest project.”
Essentially, the research consisted of hiring students to post outside of various hot spots, which provided its own challenges. The team had to hire quickly, and because of the risk participants were taking, the team was limited to hiring people willing to go outside. Next, the team assembled a group meeting over Zoom to work out project logistics. Posting in the field for 10 to 20 hours a week between March 22 and May 19, they were mostly undergraduate and graduate students. Despite the many roadblocks, enough data was collected across the city for their work to take shape.
The statistics showed changes in people’s behaviors as the weeks went by. Both Kirchner and Laefer mentioned changes in mask wearing as weather changed. For example, they noticed a spike in mask usage the first couple of weeks of May when mask wearing became widely recommended—but then a drop in mask wearing soon after the weather changed for the better.
What does their research mean for the future? “The program is about data collection and getting it out to the community,” Laefer says. The goal is to try to collate the findings into information the public, officials, and others can use. Currently, they’re in the final stretch of cleaning up the data and making it usable and accessible to everyone. “We’ve paired it with some other data sets that we think people would be very interested in using,” Laefer explains. Kirchner hopes that this crisis helps prioritize public health research: “Without funding, research like this isn’t possible.”