TABLE OF CONTENTS Multivariate data sets are readily available in many industries such as manufacturing, service, and business administration. For example, in an automobile body assembly operation, the optical coordinate measurement machine can simultaneously measure dozens of dimensional locations on each automobile in the assembly line. In a process industry, many process variables such as temperature and flow rate, are constantly monitored by sensors and enormous amounts of multivariate data are generated. On the other hand, multivariate statistical methods are very powerful analytical methods, especially with the development of computing power. Multivariate statistical methods are able to analyze huge amounts of data and extract valuable information. However, compared to other popular quality methods, such as design of experiments (DOE), linear regression, and response surface methods, multivariate statistical methods are much less known to quality professionals. One of the major concerns in applying multivariate statistical methods is the difficulty involved in communicating multivariate data and analysis results to nonstatisticians. We believe that graphical multivariate data display and visualization are excellent tools to bridge the communication gap between multivariate statistical methods and their end users. To understand a large volume of multivariate data and analysis results and to make sense out of them is a real challenge for end users or subject matter experts. However, translating data back to graphics or images relating to the end user s discipline will usually be of great help.
This chapter will describe several techniques to display multivariate data. Section 2.2 will discuss the use of graphical templates...
