Bio

Research

Dr. Hoskovec used high performance computing to compare and evaluate contemporary methods for estimating the association between exposure to multipollutant mixtures and health outcomes. To further research on personal environmental exposures, Dr. Hoskovec developed a new method for analyzing multiple multivariate time series with missing data and applied the method to an analysis of personal exposure among multiple people. The method identifies shared states of exposure and captures patterns in exposure states to improve inference on temporal and spatial sources of pollution. Motivated by the joint health risks of air pollution and the COVID-19 virus, she developed a new method to study the relationship between exposure to air pollution and COVID-19 case severity in the presence of missing COVID-19 outcome data. She created user-friendly software in the form of publicly available R packages to implement the methods she developed in her research. As a research scientist, Dr. Hoskovec's current projects include combining complex survey data to develop predictive methods for agricultural land management practices in the United States and supporting the compilation of the National Greenhouse Gas Inventory project.

Consulting

Dr. Hoskovec has statistical consulting experience in both academic and private industry settings. She provides statistical support for scientific studies in ecology, epidemiology, environmental health, occupational health, mechanical engineering, and more. Her roles include study design, model selection, data analysis, data visualization, statistical inference, report writing, and client communication.

Teaching

Dr. Hoskovec has taught at the University level since 2016. Her teaching experience includes introductory courses in statistics, computer science, and programming.