Exploring proportions: Comparative visualization of categorical data

This poster describes an approach to facilitate comparisons in multi-dimensional categorical data. The key idea is to represent over- or under-proportional relationships explicitly. On an overview level, the visualization of various measures conveys pair-wise relationships between categorical dimensions. For more details, interaction supports to relate a single category to all categories of multiple dimensions. We discuss methods for representing relationships and visualization-driven strategies for ordering dimensions and categories, and we illustrate the approach by means of data from a social survey.