Carrots are good for vision, models are good for visualising discrete data
Details
Join us and the SSA NSW branch for this talk by Dr Gordana Popovic (UNSW). You can come to University of Sydney in person (location: F10A.01.024.Law Annex Lecture Theatre 024). Or join via zoom.
Data sometimes have artefacts which obscure the patterns we are interested in visualising. You may have come across this when trying to interpret a residual plot from a logistic regression. For discrete data, these are not really artefacts, they are properties of the distribution of the data, like mean-variance relationships and lots of zeros. Generalised linear models and their extensions model the distribution of the data, including these properties, and so are a good way to remove them, to better see the patterns we are actually interested in. I will talk though some concepts and lots of examples of how we can use models to get the best out of our visualisations of discrete multivariate data in R.
Biography
Gordana is a statistical consultant and research fellow at UNSW Sydney, with research interests in modelling and visualising multivariate discrete data in ecology. In her spare time, Gordana can be found hiking or ocean swimming (poorly but enthusiastically).