Podcasts SDS 490: Say No to Pie Charts

SDS 490: Say No to Pie Charts

Podcast Guest: Jon Krohn

July 20, 2021

Subscribe on Website, Apple Podcasts, Spotify, Stitcher Radio or TuneIn

Welcome back to the FiveMinuteFriday episode of the SuperDataScience Podcast!

This week I discuss why not to use circular visualizations.

 

Circles are attractive. But displaying data as a circle is a bad idea. The pie chart and donut chart look nice but as a graphic meant to display information clearly this is the wrong approach. The circular reputation of data has a psychological issue for us: it’s hard for us to determine by sight alone which portions of a circle are larger than others, especially if they’re close in size.
So, if you’re trying to make it easy to distinguish the magnitude of data, don’t use a circular chart. Use a bar chart.
ITEMS MENTIONED IN THIS PODCAST:
DID YOU ENJOY THE PODCAST?

Podcast Transcript

(00:05):
This is FiveMinuteFriday on Not Using Pie Charts.

(00:19):
I get it. Circles are attractive. But, displaying data as a circle is rarely a good idea.
(00:28):
Case in point: Pie charts, and their close cousin the donut chart, they can look aesthetically pleasing if you put them on a document, particularly on a slide. But, if the point of a graphic is to display information clearly as opposed to as beautifully as possible, then a pie chart or donut chart is rarely going to accomplish what you’re looking for.
(00:50):
The problem with circular representations of data is that it is difficult for our eye to distinguish which slices of the circle are larger than others, particularly if the slices are similarly sized and/or distant from each other in the plot.
(01:07):
So, if you’re trying to make it easy to distinguish the relative magnitude of buckets of data, don’t use a pie chart! In the vast majority of circumstances, a bar chart — which makes it extremely easy at a glance to compare the relative height of categories — should be your go-to plot type.
(01:26):
If you’d like more on this, as well as examples of rare circumstances where you may actually want a circular representation, check out the article ‘Why use a radial data visualization?’ which is available at post at ObservableHQ.com.
(01:41):
All right, that’s it for today’s episode. Keep on rockin’ out there, folks, and catch you on another round of the SuperDataScience show very soon. 
Show All

Share on

Related Podcasts

August 19, 2025

Michelle Yi
  • 66 MINS

SDS 915: How to Jailbreak LLMs (and How to Prevent It), with Michelle Yi

August 15, 2025

Oz Katz
  • Artificial Intelligence, Data Science
  • 24 MINS

SDS 914: Data Lakes 101 (and Why They’re Key for AI Models), with Oz Katz

August 12, 2025

Julien Launay
  • Artificial Intelligence, Data Science, Python
  • 72 MINS

SDS 913: LLM Pre-Training and Post-Training 101, with Julien Launay