Martin Krzywinski / Canada's Michael Smith Genome Sciences Centre / mkweb.bcgsc.ca Martin Krzywinski / Canada's Michael Smith Genome Sciences Centre / mkweb.bcgsc.ca - contact me Martin Krzywinski / Canada's Michael Smith Genome Sciences Centre / mkweb.bcgsc.ca on Twitter Martin Krzywinski / Canada's Michael Smith Genome Sciences Centre / mkweb.bcgsc.ca - Lumondo Photography Martin Krzywinski / Canada's Michael Smith Genome Sciences Centre / mkweb.bcgsc.ca - Pi Art Martin Krzywinski / Canada's Michael Smith Genome Sciences Centre / mkweb.bcgsc.ca - Hilbertonians - Creatures on the Hilbert CurveMartin Krzywinski / Canada's Michael Smith Genome Sciences Centre / mkweb.bcgsc.ca - Pi Day 2020 - Piku
What do the trees know.Lalehsway, sway, swaymore quotes

blue: color


Scientific graphical abstracts — design guidelines


visualization + design
If you are interested in color, explore my other color tools, Brewer palettes resources, color blindness palettes and math and an exhausting list of 10,000 color names for all those times you couldn't distinguish between tan hide, sea buckthorn, orange peel, west side, sunshade, california and pizzaz.

Designing for Color Blindess

Color choices and transformations for deuteranopia and other afflictions

Here, I help you understand color blindness and describe a process by which you can make good color choices when designing for accessibility. You can also delve into the mathematics behind the color blindness simulations.

Different color blindness simulations don't all agree on the luminance of the simulated color. See methods for details.

Download all palettes
plain text, v11 20 May 2020
Download slides
PDF, v11 20 May 2020

In this section, I cover how to make good color choices when considering audiences with color blindness.

With the exception of the 8-color palette, all palettes have been created using a process (read below) that tries to maintain perceptual luminance uniformity in color blind space.

conservative 8-color palettes for color blindness

This 8-color palette is adapted from Nature Method's Points of View: Color blindness by Bang Wong. Note that in that original source the RGB values listed in the table did not exactly correspond to the RGB swatches—probably an RGB vs CMYK conversion mixup.

This palette is suitable for categorical color encoding—the colors do not, as a whole, have a natural order and none is substantially more salient than another.

You can download these colors as plain text list of HEX and RGB values.

An 8-color palette for color blindness. / Martin Krzywinski @MKrzywinski mkweb.bcgsc.ca
An 8-color palette for color blindness, adapted from Wong, B. (2011) Nature Methods 8:441. (zoom, PDF, plain text)

For more tips about designing with color blindness in mind, see Color Universal Design (CUD) — How to make figures and presentations that are friendly to people with color blindess.

using color equivalencies

To people with color blindness, some colors appear the same. This equivalence can be used to identify colors that are distinct to those with normal as well as to those with color blindness.

For a given RGB color we can simulate how it would appear to someone with color blindess and identify groups of RGB colors that appear indistinguishable in color blindness.

Color equivalencies in color blindness for protanopia, deuteranopia and tritanopia. / Martin Krzywinski @MKrzywinski mkweb.bcgsc.ca
Many colors appear the same to people with color blindness. This chart shows the colors that are identical (rows) to those with protanopia, deuteranopia (most common) and tritanopia (zoom)

These equivalencies can be used to construct color palettes—lists of colors that are distinguishable to deuteranopes and those with normal vision.

Since deuteranopia is the most common, this is the condition that I use for color selection.

The exact luminance (perceived brightness) of the simulated color varies depending on the color blindness algorithm. Each row in the squares above should look identical using any color blindness simulation (e.g. Color Oracle, Photoshop, etc) but brightness of the rows may be slightly different than shown here.

12-color palettes for color blindness

This palette maps four colors onto each of the two color dimensions in deuteranopes and four onto greyscale. This palette is very useful for designing transit and subway maps.

Color names are playful selections from my list of 10,000 color names.

You can download these colors as plain text list of HEX and RGB values.

A 12-color palette for color blindness. / Martin Krzywinski @MKrzywinski mkweb.bcgsc.ca
A 12 color palette for color blindness that maps onto each of the color dimensions in deuteranopes. Within each set of four, colors also have reasonably similar greyscale tones. Inset swatches show color alternatives that are indistinguishable from the main swatch for deuternopes. (zoom, PDF, plain text)

15-color palettes for color blindness

You can download these colors as plain text list of HEX and RGB values.

A 15-color palette for color blindness. / Martin Krzywinski @MKrzywinski mkweb.bcgsc.ca
A 15-color palette for color blindness. (zoom, PDF, plain text)

24-color palettes for color blindness

Even more color choices for color blindess, including colors that map onto greys. For these, I don't have RGB/HEX values handy.

You can download these colors as plain text list of HEX and RGB values.

15-color palettes for color blindness / Martin Krzywinski @MKrzywinski mkweb.bcgsc.ca
15-color palettes designed for each of the three types of color blindness: deuteranopia, protanopia and tritanopia. Palettes are shown as they appear to someone with normal vision as well as to someone affected with each of the three types of color blindness. Each palette contains three groups of swatches, matching to two of the color channels and greys. Within each group colors in the same row map onto the same color. (zoom, PDF)

the last word on color palettes for color blindness

You can create your own color palettes using the figure below.

For a given color blindness type (e.g. deuteranopia) and channel (e.g. blue), the rows represent reasonably uniform steps in LCH luminance of the simulated color and a rich (high chroma) simulation at that luminance.

The last word on color palettes for color blidness. / Martin Krzywinski @MKrzywinski mkweb.bcgsc.ca
Color ramps of 16–19 colors for each color channel for each color blidness type. Color ramps show RGB colors and their color blindness simulation grouped by channels (e.g. greys, blues, yellow). Within a channel, colors are sorted in increasing and roughly equal steps of LCH luminance of the simulated color. At a given luminance, the RGB color whose simulation has the highest chroma is used. (zoom, PDF)

VIEW ALL

news + thoughts

Graphical Abstract Design Guidelines

Fri 13-11-2020

Clear, concise, legible and compelling.

Making a scientific graphical abstract? Refer to my practical design guidelines and redesign examples to improve organization, design and clarity of your graphical abstracts.

Martin Krzywinski @MKrzywinski mkweb.bcgsc.ca
Graphical Abstract Design Guidelines — Clear, concise, legible and compelling.

"This data might give you a migrane"

Tue 06-10-2020

An in-depth look at my process of reacting to a bad figure — how I design a poster and tell data stories.

Martin Krzywinski @MKrzywinski mkweb.bcgsc.ca
A poster of high BMI and obesity prevalence for 185 countries.

He said, he said — a word analysis of the 2020 Presidential Debates

Thu 01-10-2020

Building on the method I used to analyze the 2008, 2012 and 2016 U.S. Presidential and Vice Presidential debates, I explore word usagein the 2020 Debates between Donald Trump and Joe Biden.

Martin Krzywinski @MKrzywinski mkweb.bcgsc.ca
Analysis of word usage by parts of speech for Trump and Biden reveals insight into each candidate.

Points of Significance celebrates 50th column

Mon 24-08-2020

We are celebrating the publication of our 50th column!

To all our coauthors — thank you and see you in the next column!

Martin Krzywinski @MKrzywinski mkweb.bcgsc.ca
Nature Methods Points of Significance: Celebrating 50 columns of clear explanations of statistics. (read)

Uncertainty and the management of epidemics

Mon 24-08-2020

When modelling epidemics, some uncertainties matter more than others.

Public health policy is always hampered by uncertainty. During a novel outbreak, nearly everything will be uncertain: the mode of transmission, the duration and population variability of latency, infection and protective immunity and, critically, whether the outbreak will fade out or turn into a major epidemic.

The uncertainty may be structural (which model?), parametric (what is `R_0`?), and/or operational (how well do masks work?).

This month, we continue our exploration of epidemiological models and look at how uncertainty affects forecasts of disease dynamics and optimization of intervention strategies.

Martin Krzywinski @MKrzywinski mkweb.bcgsc.ca
Nature Methods Points of Significance column: Uncertainty and the management of epidemics. (read)

We show how the impact of the uncertainty on any choice in strategy can be expressed using the Expected Value of Perfect Information (EVPI), which is the potential improvement in outcomes that could be obtained if the uncertainty is resolved before making a decision on the intervention strategy. In other words, by how much could we potentially increase effectiveness of our choice (e.g. lowering total disease burden) if we knew which model best reflects reality?

This column has an interactive supplemental component (download code) that allows you to explore the impact of uncertainty in `R_0` and immunity duration on timing and size of epidemic waves and the total burden of the outbreak and calculate EVPI for various outbreak models and scenarios.

Martin Krzywinski @MKrzywinski mkweb.bcgsc.ca
Nature Methods Points of Significance column: Uncertainty and the management of epidemics. (Interactive supplemental materials)

Bjørnstad, O.N., Shea, K., Krzywinski, M. & Altman, N. (2020) Points of significance: Uncertainty and the management of epidemics. Nature Methods 17.

Background reading

Bjørnstad, O.N., Shea, K., Krzywinski, M. & Altman, N. (2020) Points of significance: Modeling infectious epidemics. Nature Methods 17:455–456.

Bjørnstad, O.N., Shea, K., Krzywinski, M. & Altman, N. (2020) Points of significance: The SEIRS model for infectious disease dynamics. Nature Methods 17:557–558.