Toward measuring visualization insight

Measurement of visualization insight enables the direct comparison of visualization design alternatives as well as comparison of insight goal. The important characteristics of insight include complexity, depth, qualitativeness, and unpredictability. The insight is deeply embedded in the data domain, connecting the data to existing domain knowledge and providing its relevant meaning. There are various methods of visualization evaluation including empirical methods such as controlled experiments, usability testing, longitudinal studies, and analytical methods such as heuristic evaluation and cognitive walkthroughs. Future steps should be taken with a rigorous and comprehensive approach to compare these methods, and uncontrolled evaluation methods should be adapted to better gauge insight. These new methods can provide better measures of visualization insight and ultimately determine the purpose of visualizations.