Building a new geom in ggplot2

Hello Ladies! We are excited to host our first in person event this year at Eliiza. Looking forward to seeing you all!

Dr. Sayani Gupta is a statistician who recently completed her PhD from the Department of Econometrics and Business Statistics, Monash University on the visualisation and analysis of probability distributions of large temporal data. Her research interests are data visualisation, computational statistics, time series analysis, and forecasting. She is interested in connecting various business problems to different exploratory data analysis and statistical modeling opportunities, and almost always uses R to do that.

In this talk, Sayani will walk you through how to build new plotting symbols that are not implemented in ggplot2 yet. This enables users to customise and integrate new graphical elements with existing ggplot2 visualisations, resulting in a broad range of informative graphics. The idea is to illustrate it with a few examples, but mainly demonstrate it through the recently developed R package gghdr. The package provides a framework for visualising Highest Density Regions (HDR) using the ggplot2 graphics syntax. The method of summarizing a distribution using HDR is useful for plotting multimodal distributions, but unlike box plots, density plots, or violin plots, there is no way to draw HDRs in a ggplot2 framework. We will see how gghdr extends the functionality of ggplot2 by calling new stat_* and geom_* functions, where the stat_* functions are inherited from the R package hdrcde and the geom_* functions are built to produce the new plotting symbols.