Visual Six Sigma: Making Data Analysis Lean

5.4. Uncovering Relationships

Jane is excited about moving on to the task of identifying Hot Xs. But first, she takes a little time to think through how she will structure the analysis of her data. She thinks back to the Visual Six Sigma Data Analysis Process (Exhibit 3.29) and observes that she is now at the Uncover Relationships step, which, if necessary, is followed by the Model Relationships step. Referring to the Visual Six Sigma Roadmap (Exhibit 3.30), she decides that she will:

• Use Distribution and dynamic visualization to better understand potential relationships among her variables.

• Plot her variables two at a time to determine if any of the Xs might have an effect on the Ys.

• Use multivariate visualization techniques to explore higher-dimensional relationships among the variables.

If Jane finds evidence of relationships using these exploratory techniques, and if she can corroborate this evidence with contextual knowledge, she could stop her analysis here. She would then proceed directly to proposing and verifying improvement actions as part of the Revise Knowledge step. This presumes that she can substantiate that her findings are indicative of real phenomena. However, since her data set is not very large, Jane decides that for good measure she will follow up with the Model Relationships step to try to confirm the hypotheses that she obtains from data exploration.

5.4.1. Dynamic Visualization of Variables Using Distribution

She begins to explore which Xs might be influencing ...