Evaluating spatial- and temporal-oriented multi-dimensional visualization techniques

TL;DR: The efficacy of function-based, data-driven, spatial-oriented, and temporaloriented visualization techniques are discussed based upon extensive review and readers can find practical implications for both research and instructional practices.

Abstract: It is commonly believed that visualization tools can help researchers unveil hidden patterns and relationships among variables, and also can help teachers and speakers present abstract statistical concepts and complicated data structures in a concrete manner. However, higher-dimension visualization techniques, such as those depicting more than three dimensions, can be confusing and even misleading, especially when human-instrument interface and cognitive issues are under-applied. Furthermore, statisticians, like other humans, are vulnerable to visual illusions when viewing statistical graphs (Wilkinson, 1993). Jacoby (1991, 1998) asserts that multiple-dimension is not a mathematical problem, but remains a challenge to data visualization. From the standpoint of human perception and understanding, the potentially extreme multidimensionality of multivariate data presents serious difficulties due to many cognitive limitations, and is Visualization tools are said to be helpful for researchers to unveil hidden patterns and relationships among variables, and also for teachers to present abstract statistical concepts and complicated data structures in a concrete manner. However, higherdimension visualization techniques can be confusing and even misleading, especially when human-instrument interface and cognitive issues are under-applied. In this article, the efficacy of function-based, data-driven, spatial-oriented, and temporaloriented visualization techniques are discussed based upon extensive review. Readers can find practical implications for both research and instructional practices. For research purposes, the spatial-based graphs, such as Trellis displays in S-Plus, are preferable over the temporal-based displays, such as the 3D animated plot in SAS/Insight. For teaching purposes, the temporal-based displays, such as the 3D animation plot in Maple, seem to have advantages over the spatial-based graphs, such as the 3D triangular coordinate plot in SyStat. Page 1 of 16 Evaluating spatialand temporal-oriented multi-dimensional visualization techniques. Yu, Chong Ho & Shaw...