Tech Update Statistical data plots thicken with 3-D VRML

By Chris Jones

Publication Date: August 19, 1996 (Vol. 18, Issue 34)

As Java and Virtual Reality Modeling Language (VRML) continue to emerge as graphical development environments, statistical software vendors are seeing new possibilities for data visualization that will find their way into applications early next year. (See "VRML comes of age," Aug. 5, page 1.)

SAS Institute Inc., SPSS Inc., and Lucent Technologies Bell Laboratories are developing graphics technology that will enable users to create interactive models that have an improved capability for revealing patterns and trends in data.

SAS plans to introduce VRML objects early next year that work with their statistics applications to create interactive 3-D graphics linked to data sets. A 3-D histogram, for example, could be built with each surface color-coded to represent different response values that could be seen while a user was rotating the model.

Analysts said that statistical data visualization applications could be important business tools but only if they are easy enough to use.

"These [statistical software] developers have wanted to move their technologies into the commercial mainstream for some time in hopes of getting the decision support people to use them," said Clare Gillan, an analyst at International Data Corp., in Framingham, Mass.

Nancy Dobrozdravic, director of product planning at SPSS, said their applications now have flat 3-D scatterplot displays where users are able to rotate models and identify points. She added that SPSS has recently added HTML and JPEG support for viewing the models over the Internet. SPSS is focused on providing more distributed access to data, and it is looking at VRML and Java "to give people ways to build graphs that are more intuitive and interactive," Dobrozdravic said.

John Chambers, a researcher at Bell Laboratories, said they are using Java to give S, their statistical programming language, a graphical face and provide interactive views of data. A Java application might be built, for example, to map data and do summary plots showing patterns of defects that occur in manufacturing lines, Chambers said.