color blindness – KATHRYN MILLER

Simulation of the normal (above) and dichromatic (below) perception of red and green apples — Source: Alex Wade

Yes, color matters!

There are many different ways to present your data and information and it’s important to remember the role color plays in the interpretation of results. Many programs apply default color settings to your figures, but those colors may not be the best choice. Color vision deficiency (CVD) affects a significant percent of the population, with red-green color vision defects being the most common, though there are various types of CVD or color blindness. Contrast, saturation, and hue are important options to consider when choosing colors for your figures. In the left image, there is a depiction of the normal (above) and dichromatic (below) perception of red and green apples. In the image below, we can see how specific colors look to those with CVD.

Colors optimized for color-blind individuals

Screenshot of Figure 2 from the Nature article linked in the image. It's a table with 6 columns. 1st column: color; 2nd column: color name; 3rd column: RGB (1-255); 4th column: CMYK (%); 5th column: P; 6 column: D. Each row has a color block, name of the color, how that color is represented in RGB and CMYK, and how people with protanopia and deuteranopia see that color. — P and D indicate simulated colors as seen by individuals with protanopia and deuteranopia, respectively.
Source: Wong, B. Points of view: Color blindness. *Nat Methods* 8, 441 (2011). https://doi.org/10.1038/nmeth.1618

Data can be misinterpreted without careful color choices. Many studies (examples: 1, 2) have shown how color can lead to data misinterpretation, even for those without color vision deficiencies.

Tools you should use when creating visualizations

There are resources available to assist you in making your visualizations more accessible. Authors of a 2018 PLOS One article developed a Python module that creates a new CVD-optimized colormap. PlotTwist, a web app for plotting and annotating continuous data, also has CVD-friendly palettes. Tableau has tips and palettes for creating graphics. A 2019 Materials & Design article discusses how to improve data visualizations in the physical sciences (with a focus on materials science), including choosing the best color map. TPGi’s free color contrast checker tool has a CVD simulator. Publishers, such as ACM, are advising authors to submit visualizations that can be interpreted in greyscale, as well as providing CVD-friendly color palette tools.

The next time you create a visualization or graphic – remember: there are more choices than “Roy G. Biv“, and the vendor defaults are not always the best option.