Friday, July 3, 2015

"Don't use jet..."

Jet is the worst colormap, except for all the others.

No, Winston Churchill was not talking about colormaps. He had more important things on his mind.

Starting from Jake VanderPlas' "How bad is your colormap" and this haiku tweet you can consider the matplotlib colormaps (and references therein) and the issues of luminance. For example, in the "jet" plot (default for imshow in matplotlib, and MatLab by the way) you can see artifacts. The bright blue line is meaningless. It just comes from a poor cmap definition.

But wait, there's more! Don't miss Climate Lab Book's End of the Rainbow, If We Assume's Planck vs WMAP Cosmic Ray Background OMG and this recent Scrap rainbow colour scales article in Nature!

These are some plots of the simulation in my "Relax Laplace!" post (in the works).

I also promise to add a post with some GOOD alternative maps.

Scroll down to the second (Black and White) version to understand the artifacts better.  Keep scrolling to see other gradient patterns.

The Seaborn project is worth reading about also.


x     = np.linspace(-1, 1, 200)
X, Y  = np.meshgrid(x, x)
theta = np.arctan2(Y, X)

theta[theta<0] += 2. * np.pi

No comments:

Post a Comment

Note: Only a member of this blog may post a comment.