One University of Minnesota doctoral student stood out at a national flu forecasting competition this summer.
Each year the U.S. Centers for Disease Control holds an influenza forecasting competition where researchers seek to predict the virus’s severity as well as the onset of flu season. University public health Ph.D. student Yang Liu, who placed fourth, said her statistical model stood out because it used environmental data, which influences public health.
“I used environmental data because I think temperature variations are crucial to the health of our respiratory system,” Liu said, explaining that she included information like temperature records in her model.
During the competition, which started in fall 2016 and continued through May 2017, Liu competed against 20 other teams from across the country. She said she received feedback from her colleagues Joseph Servadio and Maria Sundaram, both environmental health doctoral candidates at the University, throughout the process.
“The influenza itself is always changing. It’s quite a challenging pathogen to predict, and famously so. People often say, ‘If you’ve seen one influenza season, then you’ve seen one influenza season,’” Sundaram said.
Forecasting can help inform the public about the virus’s potential impact ahead of flu season, which could allow people to better prepare by getting vaccinated, for example, said Matthew Biggerstaff, an epidemiologist at the CDC that helps coordinate the annual competition.
“Being able to measure the accuracy of forecast year to year is important. I think it kind of helps people understand where we are … and if we are seeing improvements,” Biggerstaff said.
Liu said she enjoys that the competition is a space for students to come up with creative ways to forecast flu seasons.
“What I really like about this competition is it’s a community of disease modelers with [somewhat] different backgrounds. We have computer scientists, mathematicians, physicists and I’m an environmental health researcher,” she said.