Appendix A. Reference Palettes#

For many applications, perceptually uniform colormaps are the best choice; that is, colormaps in which equal steps in data are perceived as equal steps in color space. The human brain perceives changes in lightness better than changes in hue, for example. Therefore, colormaps with monotonically increasing lightness are better understood by viewers.

Colormaps are typically categorized by their function:

Sequential colormaps: Lightness and saturation change gradually, usually using a single hue; suitable for data with ordered information.
Diverging colormaps: Lightness and saturation change for two different colors, meeting at an unsaturated color in the middle; suitable when the data has a critical midpoint, such as topographic maps or data deviating around zero.
Cyclic colormaps: Lightness varies for two different colors, meeting in the middle and unsaturated at both ends; suitable for values that cycle at the endpoints, such as phase angles, wind direction, or time of day.
Qualitative (discrete) colormaps: Used for unordered categorical variables and unordered paired variables. Ordered categorical variables use sequential colormaps.

Some miscellaneous colormaps have specific purposes for which they were created. For example, gist.earth, ocean, and terrain all seem designed for plotting terrain (green/brown) and water depth (blue). Therefore, we might expect divergence in these colormaps, but they have multiple inflections of unrelated colors, making them less friendly for other uses. cmrmap is designed to convert to uniform grayscale, though it seems to have minor inflections at the beginning. cubehelix is designed for smooth variation in both lightness and hue, but appears to have a small hump in the green hue region. turbo [1] is an improved version of jet, used for displaying differences in depth and density.

Color Vision Deficiency#

A common task in data visualization is representing data with color scales or continuous colormaps, typically in the form of heatmaps or choropleth maps. Several color scales are specifically designed with color vision deficiency in mind and are widely used in academia, including cividis, viridis, and parula. These scales combine a light-to-dark scale with a yellow-to-blue scale, making them monotonically increasing and perceptually uniform for all forms of color vision. In general, colormaps with monotonically increasing lightness are insensitive to various forms of color vision deficiency.

Normal Color Vision