The efficacy of stacked bar charts in supporting single-attribute and overall-attribute comparisons

Author
Indratmo, Indratmo
Howorko, Lee
Boedianto, Joyce Maria
Daniel, Ben
Faculty Advisor
Date
2018
Keywords
stacked bar chart , comparison task , user study , multi-attribute data , Information visualization
Abstract (summary)
Stacked bar charts are a visualization method for presenting multiple attributes of data, and many visualization tools support these charts. To assess the efficacy of stacked bar charts in supporting attribute comparison tasks, we conducted a user study to compare three types of stacked bar charts: classical, inverting, and diverging. Each chart type was used to visualize six attributes of data where half of the attributes have the characteristics of ‘lower better’ whereas the other half ‘higher better.’ Thirty participants were asked to perform two types of comparison tasks: single-attribute and overall-attribute comparisons. We measured the completion time, error rate, and perceived difficulty of the comparison tasks. The results of the study suggest that, for overall-attribute comparisons, the inverting stacked bar chart was the most effective with regards to the completion time. The results also show that performing overall-attribute comparisons using the classical and diverging stacked bar charts required more time than performing single-attribute comparisons using these charts. Participants perceived the inverting and diverging stacked bar charts as easier-to-use than the classical stacked bar chart for overall-attribute comparisons. However, for single-attribute comparisons, all chart types delivered similar performance. We discuss how these findings can inform the better design of interactive stacked bar charts and visualization tools.
Publication Information
Indratmo, et al. (2018). The efficacy of stacked bar charts in supporting single-attribute and overall-attribute comparisons. Visual Informatics. https://doi.org/10.1016/j.visinf.2018.09.002.
DOI
Notes
Item Type
Article
Language
English
Rights
Attribution-NonCommercial-NoDerivs (CC BY-NC-ND)