Modelling small area level data using Bayesian spatial models

Seminar Recording:

The rapid emergence of spatial databases has led to new challenges and opportunities in modelling large, complex, point referenced, point patterned and areal data. Bayesian spatial models are very useful and beneficial to model areal spatial data recorded at small area level, considering the influence of neighbouring areas. This seminar will discuss the state of art Bayesian spatial models in modelling small area level data using examples. It will also introduce some novel Bayesian semi-parametric approaches for modelling small area level data.

Dr Farzana Jahan is a Postdoctoral Research Fellow at QUT Centre for Data Science and Lecturer (fixed-term) in the School of Mathematical Sciences, QUT. She completed her PhD recently (August 2021) developing new insights into Bayesian models for spatial data. Her research interests include Bayesian spatial models, Bayesian models for aggregated or summary data and Bayesian Empirical likelihood among others. Currently she is involved in multiple projects with industry partners involving modelling health and environment related data. She is also working on a First Byte project looking at impacts of Bushfire on air quality and heath in Australia. She is also part of a Second Byte project team looking into some challenges of Bayesian spatial modelling of small area level aggregated data. Both these projects are funded by QUT Centre for Data Science and are targeted to ECRs. Apart from her research projects, as a postdoc of Social Systems domain of the QUT Centre for Data Science, she is involved in organising events such as Decode datathon and co-leading the initiative, Data Science for Social Good.