A multivariate EWMA chart plots multivariate exponentially weighted moving averages. Each multivariate EWMA point incorporates information from all of the previous subgroups or observations, but gives less weight to process data as they get older. A multivariate EWMA chart detects small shifts in a multivariate process.

Use a multivariate EMWA chart to determine whether or not the process mean vector for two of more variables is in control.

·    An in-control process exhibits only random variation with the control limits.

·    An out-of-control process exhibits unusual variation, which maybe due to the presence of special causes.

Multivariate EWMA charts allow you to simultaneously monitor whether two or more related variables are in control.

You are the production manager of a dinnerware manufacturing company. To ensure consistency, you monitor the dimensions of the dishes you produce. Your current project involves measuring the diameter and weight of salad plates. You collect 3 samples per day for 34 days. Because diameter and weight are correlated and you want to detect small shifts in the process, you create a multivariate EWMA chart.

Data: Dinnerware.MTW (available in the Sample Data folder).