Pie Charts Are Terrible

Let’s start this off with some honesty.  I used to love pie charts.  I thought they were great, just like the way I used to think Comic Sans was the best font ever.

But then I had some #RealTalk, and I’ve been enlightened in the error of my ways, and I want to pass on what I’ve learned to show people why pie charts aren’t the best choice for visualization.  For my day job, part of my work involves creating visualizations out of business data for our customers.  I picked up a copy of “Information Dashboard Design” a book by Stephen Few of Perceptual Edge.  If you’re at all interested in data visualization, I highly recommend his books, and on this site we attempt to use a lot of the principles in creating the visualizations we present to you.

But speaking specifically of Pie Charts, here’s why they’re a bad choice for your and your data:

It’s Hard To Do Comparisons

With a pie chart, the size of the angle determines the proportion on the data.  Everything adds up to a nice cool, crisp 100%.  But what if you want to know the exact numbers?  Well, you’re going to need data labels attached to your data, which can take up space and be cumbersome.

Look at these pie charts.  Can you tell me the exact value of each of the slices?  Can you order the colors from largest to smallest in each chart?

Sure, you can probably do it, but it will take a bit of work.

Now, let’s take a look at the same data, but with bar charts instead.  Same questions: Can you tell me the exact value of each of the slices?  Can you order the colors from largest to smallest in each chart?

Holy cow!  Wasn’t that much easier to figure out?  A bar charts makes better use of the space AND makes it easier for your to get more accuracy in understanding the numbers and doing comparisons to the other splits in your data.

Hat Tip to Wikipedia for the images.

Colors Don’t Mean Anything

The above examples uses colors to connect the data between the pie chart and the bar chart.  However, in a pie chart, you NEED color to differentiate the slices, and quite often you’ll have a legend that tells you what each color represents.

We’re going to look at some examples using data of World Population, with data provided via Google and the World Bank.

Is the color giving you any extra information?  No.  It’s only giving you a roadmap to see what color represents which region, and the colors being used don’t actually represent the regions in any way.

Now, let’s look at the same data in a Bar Chart.

One color.  You don’t need all the colors on the chart.  In fact, this now enables you to smartly use color to deliver more information into the same amount of pixel space.

We’ve been able to not only deliver the same information as the pie chart, but we’ve introduced a second piece of information to the same pixel space.

Not Much Information

Let’s look at the pie chart again and draw some conclusions from the image.

What we can conclude:

  • East Asia & South Asia are bigger than the other regions
  • North America is relatively smaller than the other regions
  • X is [bigger/smaller/about the same size] as Y

We can only conclude RELATIVE sizes.  If we want more information like exact numerics (or even approximate numerics), we need data labels attached to the charts.

Let’s go back to our bar chart:

  • East Asia has about 2.2 billion people
  • North America has about 300 million people
  • South Asia has about 1 billion more people than South America
  • North America is about half the size of South America
  • East Asia & South Asia are bigger than the other regions
  • North America is relatively smaller than the other regions
  • X has about Z amount of units
  • X has Z [more/less] units than Y
  • X is [bigger/smaller/about the same size] as Y

You can not only see comparisons, but with the addition of the axis along the side, you can start to get a grasp on the actual numerics behind the data.  It’s also a lot simpler to actually do comparisons to one another and get a good grasp on the differences between your data points.

Too Many Data Points

Let’s take a look at the World Bank data again, this time with a focus on country.

Yikes.  This is another pitfall with this type of visualization: Too many data points!  There’s no way you’re going to be able to draw any good conclusions from this chart.

So let’s take a look at the bar chart version:

Well, let’s be honest.  This might be a slight improvement, but it’s not much better because it’s hard to tell what the different countries are given that you can’t read all the labels.

So at this point, you need to think about what you’re trying to show with your visualization.  Is it relative population?  Is it something else to do with countries?  Always be thinking about the QUESTIONS you’re trying to answer, and pick the appropriate visualization for the job.

A scatter chart is a good way to throw lots and lots of data onto the screen and still effectively draw conclusions.  If you need accurate information about exact numerics in a large data set, then you might want to use a regular ol’ table, and include a column that has the percentage of that member against the total.  However, a scatter plot will help you see relationships in your data.

Let’s go back to our World Bank data.  It’s a very rich data source that looks at information across all cuts.  By taking a look at multiple metrics on a scatter chart, we can draw out some interesting data.

Here’s a chart plotting out Life Expectancy vs. GDP by country, grouped by region:

  • Size of the bubble gives a relative indication of population.  This is much better than the sizes of pie slices in a pie chart.
  • Colors actually mean something here, because each color represents a region as opposed to an individual data point.  You can see that Sub-Sarahan countries sit on the bottom left, East Asian Countries in the middle, and predominantly European, Central Asian, and North American countries on the right.
  • We’ve added in two new metrics in addition to just population.  We can see GDP Per Capita and Life Expectancy as it relates to population and can draw a correlation that the richer nations enjoy a longer life expectancy.

Now, while I may have created the visualization above myself, I can not take credit for the idea.  It came from Hans Rosling and a program on BBC Four, a clip of which I share with you below.

Try doing all that with a pie chart.

When Pie Charts Are Appropriate

Pie charts are appropriate in some cases.

When making a joke involving something circular:

When referencing actual pies, like this chart I made featuring our leftover pumpkin pie:

Conclusions

Remember that when you’re displaying information, you have a limited amount of space to get your points across, especially if you’re designing visualizations on a computer with a limited pixel space.  It’s all about increasing the amount of information you get and utilizing every pixel as efficiently as possible, therefore increasing the information utility.

Pie charts are bad for many many reasons, and if you’re designing a visualization, think about alternatives that could get your data across cleaner and more accurately.

2 thoughts on “Pie Charts Are Terrible

  1. Pingback: HUFFPOLLSTER: Arizona Republicans Back Veto Of SB1062 | Both Sides Clash

  2. Pingback: HUFFPOLLSTER: Arizona Republicans Back Veto Of SB1062 | The News On Time

Leave a Reply

Your email address will not be published. Required fields are marked *

You may use these HTML tags and attributes: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <strike> <strong>