Into the Metaverse: Let’s get over the anti-3D arguments in data viz
There is a school of thought in the Data Viz community that 3D data visualizations are to be avoided at all costs! (Often from similar quarters to anti-pie chart arguments)
The arguments usually go to:
- Poor accuracy because of 3D distortion
- This leads to poor speed of reading
- Because the human eye can not read 3D very well
These arguments may make some sense at a superficial level. But do they stand up to deeper inspection? Or even to actual empirical testing?
I was contemplating this recently as I drove along a highway in the Northeast US. On such roads, as shown in the image above, to differentiate the on-ramp merge lane from the other lanes, the lines are half the length of the normal. I am driving at about 120kph, the lines are in 3D, sloped away from me. Yet I can accurately tell which is which. Of course, I have some context in that I may have noticed the intersection. Even still, they are plain to see. The 3D distortion has not impacted here because I actually live in a 3D world, and my brain knows how to deal with this.
I presented a paper at the IEEE VIS 2019 conference showing that people read pie charts equally well or better than bar charts, debunking one of the famous myths of pies. At the same conference, in fact directly before me, Robert Kosara presented a paper which showed people successfully reading pie charts that were distorted with 3D projections. People can handle the distortion because they understand what is going on and their brains handle it.
There is a circumstance where a 3D visualization may not be ideal — where a static chart has elements that hide other values. Of course, if there is interaction available, then the user can move the obstruction out of the way.
I think that this is a vitally important discussion in the context of the Metaverse. If the best we can present in a 3D world is a flat screen with a dashboard, then we are not using the technology to its best. We need to have people be able to get up close and personal with their data — get immersed, get interactive, and get the experience of the data. All in a 3D, virtual world!