Eat Pies, Don't Share Them!

Creative data visualization strategies are key to both understanding complex patterns in our Big Data, and absolutely critical in being able to explain our insights simply.

A recent edition of my newsletter, The Marketing - Analytics Intersect (sign up here), focussed on one visualization that remains pervasive. As the response from subscribers was overwhelming, I wanted to make an exception and share with you...

Even a limited amount of exposure to data visualization wisdom will teach you that pie charts suck.

Yet, they creep into your work.

Let's rephrase that. You are amazing, and you have noticed that despite your loving efforts to the contrary, your peers still use pies.

: )

Maybe you've been unable to kill pies because a magical way to frame the uselessness of pies has been missing from your efforts. If yes, today is your lucky day.

First, for the 15 of you on this mailing list who are pie lovers... How different are the black slices between A and B?

Do you have your answer ready?

Scroll back up a tiny bit.

Look at the picture again.

Is the green slice B bigger or C bigger? Is either bigger than A?

Hard questions to answer, right?

That's why pies suck.

Here are the answers to the questions above, be prepared to be surprised...

Pies are a poor visualization choice because humans are not great at internalizing the different sections of a pie chart - unless the differences are dramatic. Technically speaking: Comparison by angle is significantly more difficult than by length.

All this is even worse when there are more than two pies, as you well noticed above.

(Source of above image: Wikipedia)

Doughnut charts, exploded pies, are not any better.

The only version of a pie-ish visualization that I have used in my work is the Sunburst chart....

I find it useful because I'm not comparing the sizes of the slices, rather using the visualization to traverse segments of behaviour.

In other applications of Sunburst, I've used it to traverses cross-session search behaviour of individual users by have the segments represent clusters similar type of keywords. It is amazingly insightful, even when I'm analyzing a portfolio of a half-a-million keywords.

We live in a world with way too much data, and one way to truly win influence with data is to visualize it. It is great for analysts - you see trends and patterns surface. It is great for leaders - you see pretty things instead of boring tables.

There are lots of amazing practical visualization strategies we can use, this is a passion of mine as evidenced here: Data Visualization Inspiration: Analysis To Insights To Action, Faster!

The post covers six visualization stories, covering Treemaps, Sunbursts, Packed Trees, Line Graphs, Streamgraphs, Bubble Charts (just barely on the right side of pie problems!), Stacked Square Charts, and of course Bar Charts.

I hope you'll have fun. And, if you have a fav, in the post above or otherwise, please do reply with your suggestion.

Bottom-line: Die pies die. :)

PS: If you loved this article, you can sign up here to receive my weekly newsletter that brings the very latest thinking from the world of Marketing and Analytics.