Brenna C. Kelly

I am a PhD student at the University of Utah studying machine learning methods to analyze the effects of air pollution on maternal health outcomes. My background is in population health, geography, and data science, and I am interested in advancing research along the environmental exposure pathway using data science:

• Are there better ways to predict or estimate an exposure?
• How should we model exposures?
• What is the effect of the exposure?
• How do we translate this into public health practice and/or policy?

Research areas
Applications. My main focus is on exposure science, or the study of human contact with environmental factors and the health effects of these exposures. Currently, I am studying mixtures of air pollutants and their effects on pregnancy outcomes. More broadly, I am interested in pollution, health and exposure inequities, and climate change.

Methods. For my dissertation, I am using a combination of statistical and deep learning methods for estimating exposure and the effects of exposures. I am interested in spatiotemporal analysis, as well as understanding (and remedying) bias and missingness. Scalable computing is essential for a lot of my work. As an epidemiologist, I’m particularly interested in causal methods and interpretable machine learning.

Current positions
I am a research assistant in the Department of Obstetrics & Gynecology and School of the Environment, Society & Sustainability. I am also spending my summer as a data science intern at Oak Ridge National Laboratory.