Justin Slater
Assistant Professor
College of Computational, Mathematical and Physical Sciences, Department of Mathematics & Statistics
Seeking academic or industry partnerships in the area(s) of
Applications of Bayesian methods, spatiotemporal statistics and geospatial data science, infectious disease epidemiology, population health
Available positions for grads/undergrads/postdoctoral fellows
Please refer to Professor Slater’s website for open positions
Research Areas
- Time Series
- spatiotemporal methods
- infectious disease models
- Epidemiology
- Biostatistics
- Bayesian statistics
Education and Employment Background
Prof. Slater earned his Ph.D. in 2023 from the Department of Statistical Sciences at the University of Toronto. Prior to this, he worked as a statistical consultant at Cytel Inc, and an analyst at the Institute for Clinical Evaluative Sciences. Justin also holds an M.Sc. in Statistics (Queen’s) and an Honours B.Sc. in Math and Statistics (Dalhousie).
Research Themes
Prof. Slater’s research interests span lie at the intersection of (bio)statistics, data science, and epidemiology.
Key areas of focus include:
- Bayesian statistics/learning – Prof. Slater is interested in prior elicitation to improve inference and computation and leveraging Bayesian hierarchical models in data-sparse contexts. In the coming years, he hopes to develop principled priors for infectious disease models and Bayesian mixture models. This core theme is intimately linked with the themes that follow.
- Spatiotemporal methods –Prof. Slater is looking to train the next generation of spatial statistics experts and geospatial data scientists. Spatial methods are vital in epidemiology, tech, agriculture, veterinary sciences and more. He has published multiple papers in the field of spatial statistics and epidemiology, primarily investigating the use of cellphone-derived mobility data in modelling epidemics.
- Infectious disease statistics – Prof. Slater develops methods for modelling and understanding infectious disease spread. His current work in this area focuses on reconstructing epidemic curves from incomplete or biased data. This involves combing count-valued state-space models for under-reported counts, and Gaussian process models in wastewater surveillance to estimate the prevalence of disease within a population. This will allow statisticians and epidemiologists to more accurately model epidemics, accounting for uncertainty from various sources.
- Modular inference – Complex statistical models often involve multiple components (modules) and data sources. Each component comes with its own set of biases, and Dr. Slater is investigating how uncertainty should be treated when the analyst “trusts” certain components of their models more than others.
- Statistical agent-based models – Infectious diseases like Hepatitis C Virus (HCV) spread via specific health behaviors that require individual (agent) level models. Dr. Slater is working with data from the Institute for Clinical and Evaluative Sciences to estimate HCV prevalence and burden in Ontario using complex Bayesian hierarchical models.
Highlights
- Banting-CANSSI Discovery Award in Biostatistics 2024