Humans are really bad at understanding color. And yet, scientific papers are so often covered in bright, eye-catching rainbow diagrams, as researchers try to communicate their findings in exciting, intuitive ways. But why – when humanity can't even agree on the color of a dress – should we think that our old friend ROY G BIV will help someone understand complex scientific ideas?
For decades, there has been a sector of the scientific community arguing that the rainbow needs to be retired. Not only is it entirely useless for colorblind people, it's not even that helpful for the rest of us – we just can't see color in a way that would let us properly interpret data.
"People like to use rainbow because it catches the eye," chemical and biological data analyst Jamie Nuñez explained to Scientific American. "But once the eye actually gets there and people are trying to figure out what’s actually going on inside of the image, that’s kind of where it falls apart."
The problems are manifold: rainbow colors are disorganized (how can you order red, blue, and yellow?), they are nonlinear (how can we tell which color is "halfway" between two others?), and, essentially, our brains just don't work that well with them – we naturally interpret bright colors as more important, regardless of where they occur on the "scale".
And as if the headache wasn't enough to convince us to change our ways, rainbow data visualizations can literally put lives at risk: a Harvard study found that doctors presented with a 3D rainbow model took longer, and were far less accurate, to diagnose heart disease than when shown a 2D model using a red-to-black scale.
Thankfully, science has once again come to the rescue. Researchers, including Nuñez, at the US Department of Energy have created a color scale using state-of-the-art mathematical modeling and optimization techniques that aims to solve our rainbow-related misery. Dubbed "cividis" – the rainbow scheme we're used to, by the way, is called "jet" – the scale runs from dark blue to light yellow, is usable by colorblind people, and, crucially, is "perceptually linear" – so the perceived changes in color really do match actual changes in data. Their results are published in the journal PLOS ONE.
So why, despite all its issues, do we keep coming back to the rainbow? The problem, cartographer and rainbow-critic Kenneth Field told Scientific American, is simply that "[p]eople love colorful maps." This is something Nuñez's team is all too aware of – a drawback of their color system, the study notes, is that "its minimal coverage of different colors... keeps cividis from being as aesthetically pleasing as [other maps]."
Unfortunately, it seems that as pretty as rainbows are, they're basically useless – or downright dangerous – as data tools. When it comes to scientific analysis, explained Field, "[r]ainbows cause more problems than they solve."
[H/T: Scientific American]