Recent software provides new tools for visualizing multivariate data that facilitate data analysis. We focus on (1) the learnability and use of visualization systems, and (2) the perceptual and cognitive processes involved in viewing visualizations. Effective visualization systems support a broad range of user tasks and abilities, are easy to learn, and provide powerful and flexible output formatting. Effective visualizations incorporate Gestalt and other perceptual and cognitive principles that encourage more rapid, automatic processing, and less slow, controlled processing.Computers have always provided powerful analytical tools and efficient means of "number crunching." Today, computers are also efficient at generating and presenting information graphically. This combination of computing and graphical power makes them almost ideal tools for combining data analysis and scientific visualization. Visualization can be thought ofas the process of using computers to display information visually for purposes of analysis (Hummel, 1994), using a combination of points, lines, coordinate systems, numbers, symbols, words, shading, and color to represent measured quantities (Tufte, 1983). In essence, visualizations can serve as graphical descriptions and summaries ofnumerical information. Although it is often treated as a separate topic in and of itself, visualization has the same purpose as mathematical and statistical procedures in data analysis-namely, to make apparent any trends and patterns in the data that indicate underlying relationships among variables. Data analysis is therefore a continuum of activity ranging from exploratory work to the presentation of results.In this paper, we first discuss the value of visualization using both simple (consisting offew variables) and complex (multivariate) data sets. Several examples generated from currently available software are presented. Despite the fact that visualization can add value to data analysis and that the software to generate visualizations is commercially available, there seems to be limited use of vi- sualization techniques as a regular component of data analysis in the social and behavioral sciences. The latter part of this paper presents a human-factors perspective to account for the limited use of visualization, focusing first on the factors relevant to the use ofcomputer-based systems to generate visualizations, and then on the cognitive and perceptual factors that pertain to the comprehension and interpretation of visual views of data.
THE VALUE OF VISUALIZATIONVisualization can aid in the theoretical interpretation of data by guiding the process of math modeling. For example, a Pearson correlation coefficient is often used to quantitatively describe a linear relationship between two variables. But it is important to view the data in order to assess the appropriateness ofthis measure. A scatterplot can easily expose a curvilinear trend in the data, or the presence ofan outlier that highly influences the Pearson coefficient. Thus, for both simple and com...