My research interests lie at the intersection of Spatial Statistics, Bayesian Statistics, Machine Learning, and Causal Inference, with applications in Epidemiology, Ecology, and Biomedical Sciences.
Embedded Files
Epidemiology
Image source: https://www.cell.com/trends/ecology-evolution/abstract/S0169-5347(05)00071-6
Developing and applying advanced spatial methods to understand disease dynamics is my major interest in this application domain. I recently developed a localized spatiotemporal random forest model to study COVID-19 dynamics across US counties, capturing how epidemiological, demographic, and environmental drivers shift across regions and periods. I also collaborated on a spatial Bayesian distributed lag non-linear model to quantify uncertainty and evaluate delayed climatic effects on Salmonella risk across New South Wales local health districts.
In addition, my interest extends into causal machine learning for estimating how treatments and interventions influence patient outcomes as circumstances evolve over the course of care. I aim to focus on developing methods that remain robust when a patient's history influences both the treatment they receive next and how they respond to it, a setting known as time-varying confounding, where standard regression and even many modern machine learning methods produce biased effect estimates.
Ecology
Image source: https://www.spatial-ecology.net/
In this application domain, my interest is in spatial methods for understanding how environmental variability shapes the distribution of hosts, vectors, and pathogens over space and time. Within this, my current work focuses on strengthening approaches that assess the reliability of model-based ecological predictions, including how far insights derived from observed data can be extended across new environmental and temporal contexts.
Biomedical Sciences
Image source: https://nanostring.com/blog/why-spatial-biology/
Developing methods for spatial omics data is my main future direction in this application domain, particularly models that capture tissue architecture and cellular interactions. I aim to explore approaches combining graph-based representations with probabilistic frameworks to characterize microenvironmental structure and multiscale patterns, with applications such as cell-state characterization and localization of biological signals.
I am also broadly interested in Bayesian approaches to medical image analysis, methods that jointly model spatial and temporal structure, quantify uncertainty, and translate raw imaging signal into physiologically meaningful parameters. This spans quantitative MRI, where such models help stabilize parameter estimation in the presence of noise, as well as broader questions about how spatial assumptions shape image-based statistical inference and how these ideas can be built into practical analysis tools.Â
Page updated
Google Sites
Report abuse