Optimizing colormaps with consideration for color vision deficiency

For those unfamiliar with the viridis color map, it is one of the most recent developments in creating sequential color maps that are inherently easier to interpret correctly, color blind friendly, and properly print to black and white if necessary. Note, by the way, that the last point is a natural consequence of the second! Getting better at making color maps more CVD friendly ends up helping everyone.

Viridis was created by Stéfan van der Walt and Nathaniel Smith as an open-source map for matplotlib. The website that kept track of its creation gives an excellent overview with more detailed information[0]. In the last couple of years it has been considered the gold standard of CVD friendly color maps. Other work in this space is Dave Green's Cubehelix and Fabio Creameri's Scientific Colour-Maps (SciCo)[1][2].

Having protanomally myself I greatly appreciate the work that has gone into all of these maps, and in my opinion viridis is a beautiful default map to use.

By freak coincidence I came across the pre-print of this paper yesterday, one day before publication. When comparing cividis and viridis head to head[3] I do see more detail and contrast in both the high and low end of the map, so it is a definite improvement for me!

More importantly, the method to improve viridis could be applied to other color maps too. On my end, yesterday's discovery lead to a cosy little twitter discussion that ended up introducing Crameri and Nuñez to each other, who had been unaware of each other's work[4]. So personally, I hope that the SciCo maps can be polished further with the insights from this paper.

[0] https://bids.github.io/colormap/

[1] https://www.mrao.cam.ac.uk/~dag/CUBEHELIX/

[2] http://www.fabiocrameri.ch/colourmaps.php

[3] https://pbs.twimg.com/media/Dji4wmnX4AEGkHA.png:orig

[4] https://twitter.com/JobvdZwan/status/1024288199785553920