Choose Language Hide Translation Bar
cyriac
Community Member

Utilizing and Leveraging Multivariate Control Charts

Cy Wegman, Founder and President, SY64

Some systems monitor hundreds of performance variables with time. Control charts are a powerful means of detecting problems in a system. JMP has developed tools to quickly scan through control charts to identify variables that are out of control. Many organizations utilize these “univariate” control charts for problem detection. There are scenarios where a multivariate control chart may be a superior way to monitor and detect problems in your system. Multivariate control charts combine many variables into one chart. This has the obvious benefit of monitoring only one chart instead of many. The key questions are: When might multivariate charts be better than univariate control charts? What information is the multivariate control chart conveying? In addition, if the multivariate chart appears advantageous, how do you use JMP to unravel the underlying issue you’ve detected? This presentation will review the basic nature of multivariate control charts, demonstrate how to use the JMP multivariate control chart platform, and identify techniques to uncover the information inside the chart. We will also discuss some of the limitations of these charts and compare other analysis options you may consider.