Detecting outliers graphically


My son Hank, ‘blogmeister’ of StatsMadeEasy, just forwarded me this cartoon from Randall Munroe’s blog xkcd. It dovetails nicely with my presentation to the ASQ Lean Six Sigma Conference this week titled “Friend or Foe? How to Use Graphical Diagnostics for Scoping Out Discrepant Data.”*


As I am wont to do, I started my talk with a humorous anecdote on the topic. This is a story on seeking something out using visual clues. It’s a matter of being on guard for something unusual like Nassim Taleb’s Black Swan but being a Minnesotan my thinking differs a bit on what’s distinctive.

“A fellow borrowed a neighbor’s fishing car** to sneak in some last-minute anniversary gift shopping at Mall of America (MOA) without his wife knowing. After wandering the vast hallways of MOA for a long period of time, he found just the thing. However, by then he’d forgotten where the car was and even what it looked like. All he could think of to tell the security staff was that the front right tire did not have a hub cap. After a thorough search the car was located. However, it would have been easier for the MOA staff if the fellow had thought to mention that the car had a red canoe strapped to the roof! “

*For the basis of my talk, see this manuscript.
**Fyi Minnesotans generally keep two spare automobiles — a ‘winter beater’ that stands up to all the snow and road slop, plus an old ‘fishing car’ that can be kept at the ready with rods and all (and allowed to get stinky with fish and bait).

  1. No comments yet.

You must be logged in to post a comment.

%d bloggers like this: