Expressing Uncertainty Visually

There is a long history of data graphics being used to represent uncertainty information: think of bar charts accompanied with error bars.  Whether or not symbolising statistical quantities in this way is effective at supporting uncertainty reasoning is little-discussed in social research methods primers (spoiler: it is not). In this session, I’ll take a tour through some more successful uncertainty representation techniques, drawing from examples in meteorology, human geography, medical statistics and political polling. I’ll present some principles, techniques and rules that the jobbing social scientist can use to easily improve their data analysis practice. But I’ll also, with examples, emphasise the importance of creativity and good judgement.

Roger Beecham is an Associate Professor of Geographical Analysis at the School of Geography, University of Leeds. His work typically proposes, tests and applies visual methods that help analysts to understand complex spatial processes under uncertainty. He's especially interested in how visualization relates to contemporary data analysis, and can be applied to generate and communicate evidence effectively. He is author of Visualization for Social Data Science(CRC Press).