O'Reilly logo

Visual Six Sigma: Making Data Analysis Lean by Leo Wright, Mia L. Stephens, Philip J. Ramsey, Marie A. Gaudard, Ian Cox

Stay ahead with the world's most comprehensive technology and business learning platform.

With Safari, you learn the way you learn best. Get unlimited access to videos, live online training, learning paths, books, tutorials, and more.

Start Free Trial

No credit card required

6.4. Uncovering Relationships

Sean's thinking is that a visual analysis of the data in Anodize_ColorData.jmp will give the team some insight on whether certain ranges of Thickness, L*, a*, and b* are associated with the acceptable Normal Black value of Color Rating, while other ranges are associated with the defective values Purple/Black and Smutty Black. In other words, he wants to see if good parts can be separated from bad parts based on the values of the four continuous Ys. If so, then those values would suggest specification limits that should result in good parts.

Sean realizes that this is a multivariate question. Even so, it makes sense to him to follow the Visual Six Sigma Roadmap (Exhibit 3.30), uncovering relationships by viewing the data one variable at a time, then two variables at a time, and then more than two at a time.

6.4.1. Using Distribution

To begin the process of uncovering relationships, the team obtains distribution plots for Color Rating and for each of Thickness, L*, a*, and b*. To construct these plots, Sean selects Analyze > Distribution and enters all five variables as Y, Columns (Exhibit 6.23).

Figure 6.23. Distribution Launch Dialog

When he clicks OK, the team sees the plots in Exhibit 6.24. Sean saves this script to the data table as Distribution for Five Reponses.

The distribution of Color Rating shows a proportion of good parts (Normal Black) ...

With Safari, you learn the way you learn best. Get unlimited access to videos, live online training, learning paths, books, interactive tutorials, and more.

Start Free Trial

No credit card required