Forecasting a Future: Clara Bay Uses Network Science to Predict and Prevent the Next Pandemic
As an undergraduate at UNC Chapel Hill studying applied mathematics and biology, Clara Bay had one persistent frustration: the two disciplines she loved were rarely taught in the same room. Math was taught for mathematicians; biology was taught for biologists. This frustration with the lack of integration drove her to find a PhD program that was interdisciplinary. Today, Bay is a PhD candidate in Network Science at Northeastern University, where she has built her research to sit at the intersection of mathematics, epidemiology, and public health policy, exactly the kind of interdisciplinary work she’d been searching for. 
Prior to Northeastern, Bay spent two years at the Environmental Protection Agency developing computational models of toxicology systems, in which she essentially used mathematics to explain how a given dose of a toxin affects the function of the liver and other organs. The job confirmed something important – that even when the subject matter wasn’t her passion, she loved the modeling work, “I liked the tools. It was nice to know that computational modeling was something I could see myself doing full-time.” However, she wanted to apply those tools to public health, a field of work closer to her heart. Those ambitions brought her to Northeastern’s Network Science PhD program, where she now studies to answer some of the most pressing questions in infectious disease forecasting.
While Bay divides her research into applied and theoretical, both tracks are oriented toward the same goal of giving public health decision-makers the clearest information when they need it most. On the applied side, she works closely with the CDC, which runs ongoing forecasting for influenza and COVID-19. Similar to a weather forecast, Bay and her team develop models to predict how many people will be hospitalized with the flu in each state over the coming weeks, then submit those forecasts to the CDC’s national platform. Her team also contributes to longer-term projections through the Scenario Modeling Hub, which focuses on policy-level “what-if” questions: What happens if a new COVID variant emerges? What if only a certain percentage of the population gets vaccinated? These projections have been used to inform national vaccine recommendations.
These graphs are example plots of some forecasts generated for the FluSight Forecast Hub over the past few months.
These show short-term flu hospitalization forecasts for the 2025-26 influenza season at the national (left) and Massachusetts state (right) levels. These forecasts predict the number of people hospitalized with influenza in each of the following 4 weeks.
Top row: forecasts made on December 20, 2025.
Bottom row: forecasts made on January 24, 2026.
The black line shows observed data, the teal line shows the median of the forecasts, and shaded regions represent forecast uncertainty, showing the 50% and 90% prediction intervals.
Bay’s dissertation focuses on evaluating these forecasts, so she is developing rigorous methods to quantify how accurate the predictions actually are, building ensembling tools that aggregate models from multiple modeling teams, and analyzing more than a decade of forecasting history to understand whether the field is genuinely improving, “We don’t want to build models that no one can understand. Part of what I do is make sure we’re being clear about where our models perform well, and where they don’t, and communicating our uncertainty honestly to people who rely on these forecasts.”
On the theoretical side of her work, Bay has developed a network-based model of global disease spread; one designed to answer a question that COVID made painfully relevant: how can we implement travel restrictions in ways that actually work? Historically, Bay notes, travel restrictions tend to be put in place after a disease has already spread to other locations, limiting its’ effectiveness. Her model explores whether there are more targeted ways to deploy these interventions early in an outbreak. While the model is built with respiratory illnesses in mind, it is designed to be generalizable rather than tied to any single disease, “Better understanding of how to control an emerging outbreak early on would directly improve future containment efforts for any new pathogen.”
An Unexpectedly Interdisciplinary Home
When Bay started the program, she didn’t quite know what network science was. She had looked it up but never taken a course in it. She hoped that her math foundation would be enough, and it was, but the leap still required some faith, “It was scary before I started, but once I got into the first introductory course, it clicked pretty quickly.” What she didn’t anticipate was how energizing the environment would be. Her desk neighbors work on sports analytics and the science of scientific publishing, entirely different domains from hers but connected by the same analytical toolkit, “Meeting people with similar mindsets who are applying them to completely different problems has been genuinely exciting.”
She also didn’t expect to feel so connected to the real-world applications of her work. Through collaborations with the CDC and the Scenario Modeling Hub, Bay has watched her forecasts move from her laptop into the hands of state and local public health officials making decisions about school closures, hospital staffing, and public health messaging, “You always feel like your work might just be going into the void. Seeing it used in the real world really helps contextualize what you’re doing.”
After graduating, Bay hopes to continue working in infectious disease modeling and public health, building tools with real-world policy impact – whether she stays in academia or not. For students considering Northeastern’s PhD programs, her advice is true to what her experience was: embrace the interdisciplinary nature of the University, and don’t feel like you need to arrive with everything figured out, “I didn’t come in with a specific project in mind. Being able to try different things and be proposed different projects was actually a really valuable way to discover what I was most interested in. I’d encourage incoming students not to feel like they need to have everything mapped out before they get here.”
Check out her Google Scholar page to read her publications!
Photo Credits: Nicole Samay, Clara Bay
More about Clara Bay:
Clara Bay grew up in coastal North Carolina (Wilmington), enjoying lots of time at the beach. She spent her undergraduate years at UNC Chapel Hill, where she became a college basketball fan. In her spare time, she enjoys running, reading, and spending time with her 12-year-old dog named Frankie.