Jérémie Boudreault

PhD Candidate in Environmental Health and Data Science, National Institute of Scientific Rsearch of Québec (NIRS)

Title: Estimating the health impacts of extreme heat and associated economic costs using data science approaches
Date: Friday, November 15th, 2024
Time: 1:30PM (PDT)
Location: ASB 10900

Abstract: Extreme heat has major impacts on various sectors that will be amplified by climate and demographic changes. These effects need to be accurately estimated and forecasted to reduce their consequences now and in the future. Data science approaches, including both statistical and machine/deep learning (ML) models, can be leveraged to create a more resilient future in the face of increasing extreme heat effects. This presentation had two parts. First, five ML models, including penalized regression, ensemble tree-based approaches and deep neural networks, are introduced and applied to predict the health effects of extreme heat simultaneously in multiple regions of Quebec, Canada. Models are trained to predict heat-related mortality and morbidity with several meteorological, regional and temporal predictors. Second, statistical models from the environmental epidemiology field are employed to model heat-health exposure-response functions for these same health outcomes. The resulting functions are then used to project the health burden and associated economic costs in both current and future periods, incorporating climate models and demographic projections. These two examples of data science applications for modelling heat-related health impacts open the door to a wide variety of innovations for other environmental- and climate-related hazards.