Tag Archives: bad chart

Bad 3D Pie Chart Alert! By Scientific American no less!

Well, this is a stunning piece of poor quality work by a respected magazine! I’m not sure if I can even list all of the ways in which it skews and misrepresents the data.

Let’s start with the plain fact that it’s in 3D which means the slices of the pie bear no relation to the numbers they are intended to represent. Add to that, the depth dimension (the red/brown section) which isn’t noticeable until you read the legend and by some strange chance see the extra label. Then, we’ve got pies arrows pointing in all different directions so the location of each pie actually has little relation to where they are placed on the map. Top things off with some horrible colour choices which follow no traditional conventions other than being the author’s favourite colours or matching the theme of the issue that month. Sure, I like pink, but really?

Chart at your own risk, but watch out for the pie chart nazis. Bad charts WILL be caught!

Dear @BernieMalinoff, your wish is my command

[tweetmeme source=”lovestats” only_single=false]Bernie tweets:
2.5 days of travel/meetings – need to fit 5 days into what’s left. Maybe @LoveStats can suggest a pie chart for this dilemma.


Annie responds:
I have analyzed the technical requirements of Bernie’s request and have implemented various techniques to ensure maximum clarity. I have also selected the appropriate graphical representation and applied classic infographic standards for readability and efficiency. Please leave your comments of amazement and wonderment below. Autographs to follow for those deemed important enough.

Read these too

  • So you think you have a probability sample
  • Cronbach’s Alpha, My Favourite Statistic
  • Permissible pie charts and 1 bar chart
  • Esomar Chicago: Day 1
  • Tagxedo for colorful shapeable saveable word clouds
  • Pie Charts – Our Evil Friend

    Everyone who knows me knows I hate pie charts. How could I, you ask, hate such a beautiful visualization of percentage data? Well, i just outright request that people not use pie charts because the charts are often taken far beyond their intent. All of our new and fabulous software allows even the most unskilled person to create amazing charts. And sometimes, or most times, they’re made into 3D, rotating, florescent monstrosities.
    The purpose of a pie chart is to clearly display relative percentages. The intent is to be able to quickly determine that one slice of the pie is slightly bigger or slightly smaller than another slice. When a chart has been tilted for dramatic effects, the slices become skewed. I’ve even gone so far as to run (non-probability, convenience sample) tests to see how people interpret 3D pie charts. What I found is that people are unable to determine which slice is the bigger slice! In fact, if you look at the picture here, you’ll various slices, each representing a proportion. What you don’t see, though, is that THREE of the slices in fact represent the exact same proportion. Three of them are 11%, one is 22% and one is 44%. Unless your job requires you to misrepresent data, this is certainly not the way to go.
    And that is why I hate pie charts!

    Related Articles


  • 3 Simplistic Requirements to Leverage Transparent Relationships for Synergistic Returns
  • Arriving Yesterday: A New Era of Research
  • Research, MeSearch, WeSearch, YouSearch by Dr. Seuss
  • Preaching to the Research Choir
  • How to be unfollowed and other twetiquette StatsSuck. That is all.
  • I’m smarter than all of you put together
  • %d bloggers like this: