Every data scientist should tell you that, because there is actual science that pie charts are bad visualizations. The human brain is bad at comparing angles and bad at comparing irregular areas. This means that a) you have to think to understand a pie chart and b) your understanding will be very imprecise. A good visualization is designed to take advantage of how our brains are hardwired (size, position, shape, colour) so that the user can intuitively understand what the data is without thinking, saving the thinking for what the data means.
I'd say that every time someone uses a pie chart Stephen Few spins in his grave, but thankfully he is still alive and well. He probably gets a chill go up his spine though.
Unfortunately, understanding data is only one of the things charts need to solve. Often they also need to capture the reader to be interested in the data. They also have not scare off a reader for looking complex.
I don't doubt pie charts suck at conveying data correctly, for the reasons you've given, but they may be appealing to readers nonetheless. Formulating studies to account for such preferences and impressions can be annoyingly hard -- and equally trying to account for the results and make good reports. :(
That's a fair point. I suppose I'm the viz equivalent of one of those health-food obsessed people who never eats dessert or junk food and trying to push that behavior on to everyone else 😃
I will amend my previous statement then. If you want to do any serious analysis of your data, do not use a pie chart. But if you want to throw in some pleasing empty-calories to your dashboard go ahead and add a pie chart. And it's OK to have a cheat-slide every now and again and just show a big ol' pie chart.
For further actions, you may consider blocking this person and/or reporting abuse
We're a place where coders share, stay up-to-date and grow their careers.
Every data scientist should tell you that, because there is actual science that pie charts are bad visualizations. The human brain is bad at comparing angles and bad at comparing irregular areas. This means that a) you have to think to understand a pie chart and b) your understanding will be very imprecise. A good visualization is designed to take advantage of how our brains are hardwired (size, position, shape, colour) so that the user can intuitively understand what the data is without thinking, saving the thinking for what the data means.
I'd say that every time someone uses a pie chart Stephen Few spins in his grave, but thankfully he is still alive and well. He probably gets a chill go up his spine though.
Unfortunately, understanding data is only one of the things charts need to solve. Often they also need to capture the reader to be interested in the data. They also have not scare off a reader for looking complex.
I don't doubt pie charts suck at conveying data correctly, for the reasons you've given, but they may be appealing to readers nonetheless. Formulating studies to account for such preferences and impressions can be annoyingly hard -- and equally trying to account for the results and make good reports. :(
That's a fair point. I suppose I'm the viz equivalent of one of those health-food obsessed people who never eats dessert or junk food and trying to push that behavior on to everyone else 😃
I will amend my previous statement then. If you want to do any serious analysis of your data, do not use a pie chart. But if you want to throw in some pleasing empty-calories to your dashboard go ahead and add a pie chart. And it's OK to have a cheat-slide every now and again and just show a big ol' pie chart.