Visualising data

Abstract

Keeping as much information about data as possible allows for exploration of possible structure without apriori knowledge. We will show examples of this Exploratory Data Analysis approach championed by Tukey and developed at AT&T and Bell labs. The labs and follow-up developers such as R. Gentleman and R. Ihaka 1 have made high quality software available (S,R,xgobi,ggobi) that makes even visualisation of high dimensional data highly effective. We finish the presentation with a small example of how visualisation has helped us improve a method for estimating the bias in coin tossing. Copyright © 2006 by Imperial College Press.