Stop using rainbow colourmaps

Far too often in scientific publications, the choice of colourmap for a figure has been given no thought. How do I know? Because every colour in the rainbow has been used.

For some reason rainbow colourmaps are the default in many programs, despite them being inappropriate for countless situations. I am not the first to recommend never using the rainbow/jet/hsv/etc… colourmaps, but apparently it needs repeating. If I arbritrarily pick a journal and go to the most recent articles, I am bound to come across multiple instances of these awful colour schemes.

What’s wrong with it?

There are several arguments against a rainbow (especially ‘Jet’) colourmap:

  • Changes in luminance (essentially brightness) moving between red, yellow, green, and blue affect our perception of variation within the figure. Look at the Jet colourbar below and you will find that certain parts jump out more than others. Contrast this with the grayscale colourbar. Both are meant to represent a smooth gradual change, but only the latter does.

jet_vs_grayscale

  • It is meaningless when converted to black and white. Below is the standard Jet colourbar converted to grayscale, which is how many people will see it after it’s printed in black and white.

jet_grayscale

  • It has no physical meaning. If a figure involves a colourmap, chances are that the data will be sequential (e.g., raw temperature data) or diverging (e.g, temperature above or below freezing). Examples of appropriate colourmaps for each are shown below.

diverging_vs_sequential

What to use instead

Some personal favourites of mine are hot and cold colour maps:

cold

hot

Sometimes monochromatic colour schemes are all you need. These are as simple as grayscale but may be more appealing.

monochromatic

Otherwise, check out colorbrewer for inspiration.
Most importantly, experiment.

Read more

A Self explanatory argument against a rainbow colormap: After viewing this website, ask yourself if your scientific figure should share the same colourmap

Borland and Taylor (2007) Rainbow Color Map (Still) Considered Harmful: The science behind the problems with rainbow colourmaps

Rainbow Color Map Critiques: An Overview and Annotated Bibliography
Why Matlab finally changed their default colourmap

Advertisements

Author: hugke729

PhD student in physical oceanography

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s