It is interesting how we researchers sometimes do plenty of research to find out what we know already. We sometimes like to complicate things too much. In the September 2012 issue of Nature Methods, two researchers, Nils Gehlenborg and Bang Wong commented that three dimensional visualization is only good for spatial data and not much for graphs. In other words they are saying: keep it simple but they have such a complicated way of saying that…. they talk about how 3D cues come from occlusion, or converging parallel lines. According to them a 3D object like that starts to occlude other data behind the bars, in addition the three dimensionality is not really beneficial and besides humans are not really good at judging lines in 3-dimensions anyway.
So instead of saying KISS – Keep it simple stupid and then STOP, they write a paper (Gehlenborg N & Wong B, Nat. Methods 9, 213 (2012) and then present a “Point of View” in another article. Vol 9 851 (2012) and have written other worthier articles every few months on the same principle and now this blog mentions them too!!
So this picture of a 3D graph is pretty much useless even though it may look very cool.