Thread by @moriah_taylor58

Having trouble picking a color palette for your #Rstats visualization? Well here's a MEGA thread about all the ways you can choose a palette! 🧵[1/22]

We'll start off with the classics. First, you can use the palettes that are already a part of R's base code (keep in mind that these palettes were not necessarily designed to be colorblind-friendly)

[2/22]

The next classic set of palettes comes from {RColorBrewer}. You can use the command display.brewer.all(colorblindFriendly = TRUE) to see which palettes are colorblind-friendly.

(A guide to this package: datanovia.com/en/blog/the-a-…)

[3/22]

Another very popular package used for visualizations is {viridis}. The great thing about these palettes is that they're ALL colorblind friendly.

(Documentation: cran.r-project.org/web/packages/v…)

[4/22]

Now onto some lesser-known packages. This one is starting to gain popularity in the community: {wesanderson}. This package has a variety of palettes inspired by Wes Anderson films which are all colorblind-friendly.

(Documentation: github.com/karthik/wesand…)

[5/22]

Another package which was inspired by the wesanderson package is {ghibli} which has palettes inspired by Studio Ghibli movies which are also all colorblind-friendly. PonyoMedium is one of my favorites!

(Documentation: ewenme.github.io/ghibli/)

[6/22]

The package {ggsci} contains color palettes inspired by colors used in scientific journals (FYI, these palettes may not be the best choice when trying to create visualizations that are colorblind/printer/photocopy-friendly)

(Documentation: cran.r-project.org/web/packages/g…)

[7/22]

Get palettes that are used by Open Analytics (link: openanalytics.eu) using the package {oaColors}

(Documentation: cran.rstudio.com/web/packages/o…)

[8/22]

The Python package 'seaborn' has some great color palettes, including a color-blind friendly one. You can load these colors in manually using this R code from GitHub: github.com/timelyportfoli…

[9/22]

None of these preset palettes tickle your fancy? Make your own! There are a variety of packages and websites that can help you do this.

(Pictured: example of a visualization I made with a custom color theme)

[10/22]

If you don't want to use HEX codes, R has 657 built-in color names. See this cheatsheet for more information: nceas.ucsb.edu/sites/default/…

[11/22]

Generate some random colors using the package {randomcoloR} where you can specify things like hue and brightness. Here are a couple examples: hue="red" and luminosity="light"

(Documentation: github.com/ronammar/rando…)

[12/22]

Making an elevation map and want it to look like the actual place it's from? Use the package {earthtones} to derive a color palette from a particular location on Earth

(Documentation: cran.r-project.org/web/packages/e…)

[13/22]

Use the package {qualpalR} which, when given a number of colors (n) and an HSL color space, provides n colors which maximize smallest pairwise color differences.

(Documentation: github.com/jolars/qualpalr)

[14/22]

Use colors from the i want hue website (link: medialab.github.io/iwanthue/) with the R package {hues}

(Documentation: github.com/johnbaums/hues)

[15/22]

Another really interesting/cool way to build a custom color palette is to generate one from an image, which you can do thanks to the package {paletteR}

(Documentation: github.com/AndreaCirilloA…)

[16/22]

Use palettes from the website COLOURlovers.com with the R package {colourlovers}

(Documentation: github.com/andrewheiss/co…)

[17/22]

You can also create a palette on the website Data Color Picker (link: learnui.design/tools/data-col…) which generates palettes for you with lots of options, including viewing the palette on a dark background, and selecting a single-hue or divergent palette.

[18/22]

REGARDLESS of which method you use, please keep the colorblind community in mind. There are numerous tools out there to test whether your color palette is colorblind-friendly, I'll provide a couple that I use.

[19/22]

My personal favorite is the Viz Palette tool (link: projects.susielu.com/viz-palette) created by Elijah Meeks and Susie Lu.

[20/22]

There is also an R package {colorblindR} that simulates colorblindness on R figures.

(Documentation: github.com/clauswilke/col…)

[21/22]

I hope this thread inspired you to branch out and use fun color palettes! Happy plotting!

[22/22]