What can possibly be worse than a 3D chart or a pie chart? A 3D pie chart. In this post, I will argue that it is actually okay to make 3D pie charts. Or, at least not as bad as some people believe and are happy to make you believe. I have never made a 3D pie chart and I doubt I ever will. I also do not find 3D pie charts aesthetically pleasing, but that is not a sufficient argument to be against 3D pie charts.
There are good reasons not to make 3D charts and not to make pie charts. I have even written posts about why you should not use pie charts and why you should not use 3D charts. I still believe there is good advice to be found in these posts, but I also like to consider the fact that it is indeed only advice and not universal rules.
My defense of 3D pie charts consists of two points. First, when communicating to an audience with a preference for 3D pie charts, 3D pie charts can be the optimal choice of data visualisation. Second, when it is not more difficult to read or understand the main takeaway because of the third dimension of the pie visualisation, there is nothing inherently wrong with this type of visualisation. In other words, if a 3D pie chart can help engage the audience without hindering the communication of the main message, a 3D pie chart is not only okay but can be the best possible visualisation.
There is some evidence that people find 3D pie charts beautiful. Burch (2015), for example, found that people had an aesthetic preference towards 3D pie charts. While there are certainly limitations related to the specific study, it is fair to conclude that some people do find 3D pie charts beautiful. When you create a data visualisation, it is not about what you find aesthetically beautiful, but about what appeals to the audience.
Next, Svalina et al. (2024) found that there is no difference in visual information processing between 2D and 3D visualisations. While this is for bar charts, I do not see any reason why this should be different for 3D pie charts. In other words, people are not more or less likely to interpret a visualisation in an incorrect manner just because it is in 3D.
If you would like to make a 3D pie chart in R, you can check out the {ggthreed} package (it is a bit old so I am not sure it is fully functional). I will most likely never make a 3D pie chart, but I am – in principle – not against such charts.