Visualization and analysis of multi-variate data: a technique for all fields

Abstract

Scientific Visualization has been driven by the need to visualize data sets that involve a large number of points and/or many dimensions or variables. However, many, or, perhaps all, fields of endeavor that use numeric data (including, finance, market research, management, manufacturing, process control, risk analysis, social services, health sciences, social sciences, physical sciences, computer science, applied mathematics, the engineering disciplines, and a host of other fields) deal with data sets of this type. Asking who can benefit from data visualization is like asking who can benefit from math or statistics. We present a hierarchical technique for visualizing truly multi-dimensional data that can be applied to any or all of these fields. The emphasis will be on visual statistical analysis of either discrete variables or continuous variables that have been sampled on, or binned to, a regular n-dimensional lattice.