Are you interested in joining Eco-Stats? Below are some project ideas - ranging from methodological to applied, depending on your interests and expertise, or of course be a mixture of both. Methodological projects will focus on developing and evaluating new methods of data analysis, and exploring their properties. Applied projects typically involve collaborating with ecologists working at UNSW and elsewhere, using modern statistical methods to address important research questions arising in ecology.
Not 100% sure the ecology focus is for you? Firstly, note that prior experience in ecology (while useful) is not required, just a willingness to learn! But secondly, problems we deal with are encountered in other disciplines also (e.g. bioinformatics, epidemiology) so skills you hone in our group enable all sorts of interesting research careers. Eco-Stats Alumni have gone on to work as biostatisticians at top universities in London, academics in Statistics and Computer Science departments, and research positions at medical institutes.
Eco-Stats researchers are based in the School of Mathematics and Statistics, but are also affiliated with the Evolution & Ecology Research Centre. As such, it is possible to do an Eco-Stats project while being enrolled through either these organisational unit.
High-throughput data (such as environmental DNA data) typically comes with information collected jointly on thousands of taxa. How can we fit models efficiently to data of this type? We will explore a number of options centered around extensions of generalised linear latent variable models that are scalable to high dimensions.
Often predictor variables (in ecology and elsewhere) are measured with error, and failing to take this into account biases estimates of the fitted model, and often, subsequent predictions. We have been developing easy-to-use algorithms for modelling such data, having initially focussed on generalised linear models for data with measurement error that is independent across observations. Important extensions include: how to extend to spatially correlated measurement error? How to generalise to handle general predictive models (beyond GLM)?
Long-term monitoring data can be used to understand how species distributions have been changing as the climate has changed. How have species been changing, and can that be linked directly to historical changes in climate? Can we use point process models, fitted to point event data, to understand how species distributions have changed over time?
