dcsimg
 

Pareto Power!

One of our Facebook followers recently wrote to us about his frustration with Pareto charts. He was confused about creating them in Minitab—especially navigating the dialogue boxes. I wanted to craft this blog post for him, but I was also thinking of an interesting way to make Pareto charts a little more fun for everyone.

Enter the Power Rangers! I grew up in the ‘90s when Power Rangers were all the rage, and—you guessed it—I was all about the Pink Ranger. I even remember dressing like the Pink Ranger for Halloween:

Pink Power Ranger

Anyway, the premise of the show is that young teens are recruited and trained to “morph” into Power Rangers. When morphed, the Rangers become powerful superheroes wearing color-coded battle suits, each possessing superhuman strength and durability, as well as a special trait and/or weapon that is unique to each Ranger (like invisibility). Their goal is to utilize their special powers, and colossal mechanical assault machines called Zords, to defeat evil forces that threaten humanity.

Power Rangers!

There’s a lot of pressure placed on the Zords to come through at the end to help the Power Rangers defeat evil forces. Occasionally a Zord suffers from a defect, such as a mechanical failure, misfire, or inability to join with another individual Zord to form a more powerful “Megazord.”

How Do You Know Which Defects Are Most Important?

Suppose the Power Rangers are involved in a quality improvement deployment. The Rangers know they will be able to save humanity more efficiently if they focus on defects that are most responsible for hampering their efforts to overcome evil forces. They want to rank the defects they encounter when using Zords so they can focus on improving the defect that will help them make the largest gains.

Pareto charts can help all quality professionals (including Power Rangers) prioritize quality problems and separate the “vital few” problems from the “trivial many.” So let's assume the Power Rangers turn to Minitab Statistical Software and the Pareto chart to help them graphically rank Zord defects from largest to smallest.

The Rangers track the number and type of Zord defects they encounter, and then record their data in a Minitab worksheet:

Minitab Worksheet

They choose Stat > Quality Tools > Pareto Chart in Minitab:

Minitab Menu Path

When presented with this dialog box, the Rangers enter Defect in the “Defects or attribute data in” field, and Count in the field for “Frequencies in,” and then they click OK:

Pareto Chart Dialogue Box

Minitab then creates this Pareto chart:

Pareto Chart of Zord Defects

How to Interpret the Pareto Chart

This Pareto chart tells the Power Rangers that 45.8% of Zord defects are attributed to pilot errors, and 36.1% are attributed to weapon misfires. Add those percentages together, and the cumulative percentage (that red line on the graph) for pilot errors and weapon misfire is 81.9%. Thus, to get the greatest improvement to Zord performance, the Rangers should focus their quality initiatives on solving pilot errors and weapon misfires.

Since the Power Rangers are the pilots and often operate the Zords, their next step might be to brainstorm the most common types of pilot errors and reasons for misfires. They might even turn to the Pareto chart again, this time to help them drill down further and prioritize the different types of defects that comprise each of the broad "Pilot Error" and "Weapon Misfire" categories they've already identified.

As Power Rangers are out to defeat evil villains with super powers, quality professionals can keep Pareto charts in their toolbox of “super powers” to defeat the enemy of quality—variability in the form of defects. Go, Go Power Rangers!

For more on Pareto charts, check out these posts:

Using a Pareto Chart: Fast Food and Identifying the Vital Few

Perils, Pitfalls, and Pareto Charts

7 Deadly Statistical Sins Even the Experts Make

Do you know how to avoid them?

Get the facts >

Comments

Name: Andy • Thursday, October 11, 2012

Indeed, good diagnosis on the "pilot error." Could it be taken a step further to see if Alpha, the Rangers robot assist, could help keep us in making an Alpha error between 2 seemingly different Pareto bars? A confidence interval based on sample size for estimation on "pilot error", indicates "pilot error" is really significant. So much for blaming the equipment!


Name: Carly Barry • Thursday, October 11, 2012

Hi Andy - thanks for your comment! It's very clever of you to make a connection between Alpha (the robot assistant) and Alpha (the error). I wish I would have thought of that! Thanks again for reading and for your follow-up analysis.


blog comments powered by Disqus