The visualization of categorical datasets is an open field of research. While a number of standard diagramming techniques exist to investigate data distributions across multiple properties, these are rarely geared to take advantage of additional data properties – either given or derived. As a result, the data display is not as expressive as it could be when incorporating these properties, and it misses out on the potential of leveraging these properties for the data's interactive exploration. In this paper, we present the visualization technique Parallel Hierarchies that is specifically tailored to take hierarchical categorizations into account. With Parallel Hierarchies, it is possible to individually adjust the desired level of detail for each categorical data property through drill-down and roll-up operations. This enables the analyst to selectively change levels of detail as the data analysis progresses and new questions arise. We illustrate the utility of Parallel Hierarchies with a demographic and a biological use case, and we report on a qualitative user study evaluating this visualization technique in an industrial scenario.
Vosough, Z., Hogräfer, M., Royer, L. A., Groh, R., & Schulz, H. J. (2018). Parallel hierarchies: A visualization for cross-tabulating hierarchical categories. Computers and Graphics (Pergamon), 76, 1–17. https://doi.org/10.1016/j.cag.2018.07.009