You analyzed your data, you made nice charts, you are ready to present your results and conclusions 👌
Wait a second, did you check if your diagrams are colorblind-friendly?
Poor color choices overall alter the efficiency of your message, and it’s even more true if your public is a colorblind person.
Color vision deficiency is not rare since it affects about 5% of humans. This represents millions of people. If you do not pay attention to your color choices, you potentially deprive a decent part of your audience from accessing your work and understanding the key points you are trying to convey.
Learn more about the biology of color vision
Since many people may not perceive the colors as you do, you may make it hard for them to read your diagrams if the colors you chose to represent your data are not distinguishable enough. At best they will have to spend extra time reading your graphs to make sure they do not wrongly interpret them; at worst they will move on quickly and ignore it since they cannot decipher them. How frustrating both for you and them!
Thankfully, there are ways to make sure that colorblind people will be able to view exhaustively the info you share, so they get to enjoy your work properly.
Go directly to the tools presentation:
• Test any color and create custom palettes with Coloring for Colorblindness
• Color choices for maps with ColorBrewer
• ColorBrewer and Viridis in R
• Coblis, to check images
• Software built-in tools
• The full-screen experience
First, contrast is important: increased difference in lightness between two colors will help to better distinguish them. When you use hues with a similar level of lightness, you increase the chance they might be indistinguishable. On the contrary, use various levels of lightness: pale colors, dark colors, and intermediate lightness levels as needed.
Moderate modification of the hue can also make the difference: you may obtain a color you would designate the same, but the change can be sufficient for others to distinguish it from another color.
For a while, my way of testing for colorblind-friendliness of my figures was to convert pictures to grayscale. It allowed me to know if the colors I chose for my plots were contrasted enough to be distinguishable by people with color-vision impairment. This method is the best way to encompass all types of color vision impairments in your color choices, including monochromatic vision. This is also useful to evaluate if your work will remain readable when printed in grayscale.
Whether for a plot, a map, or any other type of design, there are many resources to guide people with normal color vision in their color choices. Some color palettes have been designed with the aim to be adapted to color vision deficiency, and various filters or simulators calculate color rendering as viewed with an altered color vision, so that people with normal color vision can appreciate how their creations may be seen.
Below I present some of these useful resources that can help make your projects more accessible and inclusive.
► Test any color and create custom palettes with Coloring for Colorblindness
This online color-picker tool by David Nichols is one of my favorites. You can enter the color code (HEX, RGB, or HSL) of any color of your choice and check the simulated color perceived by a person experiencing protanopia, deuteranopia, or tritanopia.
You can add and compare as many colors as you want, to build custom accessible palettes that convey the color differences and contrasts you are looking for. The color palette you create is encoded in the page’s URL: you can easily save or share your color palettes simply by copying the URL.
► Color choices for maps with ColorBrewer
The cartographer and professor of geography Cynthia Brewer developed sets of colors known as the Brewer palettes, that include numbers of colorblind-friendly palettes.
The rendering of the Brewer palettes can be tested for maps with ColorBrewer 2.0 which offers the option to show only colorblind safe palettes. Similar to Coloring for Colorblindness, the page’s URL changes as you set a color scheme and the number of data classes, leaving you with the possibility to share a link to the selected color scheme.
► ColorBrewer and Viridis in R
The Brewer palettes above-mentioned have been incorporated into the R package RcolorBrewer.
To display colorblind-friendly palettes from the Brewer palettes in R:
library(RColorBrewer)
display.brewer.all(colorblindFriendly = T)
To show the colorblind-friendy palettes info:
brewer.pal.info[brewer.pal.info$colorblind==TRUE,]The viridis palette is also well known for its color vision deficiency accessibility. It was originally designed for maps in the Python package matplotlib, with the specific goal to be colorblind-friendly and better indicate color differences than most plotting default color scales.
The viridis R package includes the original viridis scheme along with six alternative palettes sharing similar characteristics.
► Coblis, to check images
With Coblis you can upload a picture of your choice and simulate any type of color vision deficiency.
Normal view of the uploaded image
Coblis simulation of deuteranopia
All the calculations are made locally on your machine, so the images you provide are not uploaded to their server.
► Software built-in tools
Some software directly includes the possibility of testing for color vision accessibility:
– In Firefox: the Accessibility Inspector in Firefox Developer Tools or the Let’s get color blind add-on to view webpages
– QGIS Preview Mode and ArcGIS Pro Color Vision Simulator
– Inkscape Color Blindness filter and GIMP Color Deficient Vision display filter. The Inkscape filter simulates color blindness (dichromacy) as well as color weakness (anomalous trichomacy).
► The full-screen experience
Color Oracle probably offers the most complete experience, with a color-blindness simulator that applies color filters to the entire screen, independently of the software currently in use.
The tools presented here do not all use the same mathematical calculations to produce simulations, so results may vary depending on the tool used. Furthermore, all these simulations are approximations and are unlikely to reflect the exact color perception of any given viewer. It is still possible that despite using these tools, some people will not be able to distinguish between the colors you chose, however, you strongly limit this eventuality.
If you really cannot adapt your color choices for a project or you want to strengthen the distinctness between the elements you show, consider using different symbols or shapes for the least contrasted elements, or even adding lettering or words.
When I conceive your graphs and illustrations, I systematically keep in mind that part of your audience may have trouble differentiating between some colors. I carefully consider the colors I pick to make my work (and so yours!) inclusive and accessible to the most people.
What about you, do you take color vision deficiencies into account when making color choices, and what are your favorite tips and tools to achieve that?
Morgane Gillard, PhD