Good design is important. Don’t believe me, maybe you’ll like this TED radio hour episode. Good design is especially important in data visualization and data communication. Colors, scales, jitter, figure type are all decisions that need to be considered. Two of my favorite palettes are
Each of these packages have additional color palettes that can be used, but these two are my favorite. The gdocs palette from
ggthemes has 20 colors, more than enough for most applications. The default
viridis color is generally used for continuous variables, however it can be used for discrete values if used with
discrete = T argument.
It is possible to use these colors outside the
ggplot2 world. Which is great, because the default colors for base R graphics are even more attrocious than
ggplot2. use the following function for each package,
viridis(n) where n is the number of colors needed.
You can use the same functions to grab colors anytime you need a color palette. Even in leaflet.
Just a quick post about making better figures. The
ggthemes has many more color palettes, all of which can be accessed with
x_pal() where x is the name of the ggtheme. For whimsical color palettes be sure to check out the
wesanderson package, a set of palettes inspired by the films of Wes Anderson. Or the
yarrr package, or learn to create your own.