Visual Six Sigma: Making Data Analysis Lean

6.9. Conclusion

Using this case study, let us review how the Visual Six Sigma Data Analysis Process aligns with the DMAIC framework, and how the Visual Six Sigma Roadmap was used to make progress quickly:

• Frame the Problem occurred in the Define phase.

• Collect Data began in the Measure phase, where the team collected data for its MSA studies and for the baseline control chart. Also, the team collected a set of historical data relating Color Rating, the team's primary, but nominal, Y, to four continuous Ys, namely Thickness, L*, a*, and b*, that were thought to provide more detailed information than Color Rating itself.

• Uncover Relationships was the goal of the Analyze phase. The team members first visualized the five Ys one at a time using Distribution, also using dynamic linking to start to explore conditional distributions. Then, they dynamically visualized the variables two at a time with a Scatterplot Matrix. Finally, they dynamically visualized the variables more than two at a time using Scatterplot 3D. From the relationships that they uncovered, they were able to define specification limits for Thickness, L*, a*, and b* that corresponded to nondefective Normal Black parts.

• Model Relationships occurred in the Analyze and Improve phases. Here, the team studied five potential Hot Xs for the four continuous Ys. A customized experiment that allowed the team to identify which Hot Xs to include in each of the four signal functions was designed and conducted. The resulting models were ...