In jiho/chroma: Parse and Manipulate Colors, Create Perceptually Correct Color Palettes

knitr::opts_chunk$set(comment="#", tidy=FALSE)

chroma

chroma is a R package for parsing and formating colors in various specifications, manipulating colors, and creating nice color scales and palettes. Much of the functionality is based on the excellent chroma.js javascript library by Gregor Aisch.

Installation

chroma is not on CRAN yet. To get it from this page, install the package devtools and

devtools::install_github("jiho/chroma")
library("chroma")

library("ggplot2")
devtools::load_all(".", quiet=TRUE)

NB: some functions in the chroma package mask functions in basic R packages (e.g. grDevices) or often used ones (e.g. ggplot2). Most of the time, they are argument-compatible with the function they mask and, if their result is different, I would argue that the one in chroma is more correct. If you want to use chroma but would rather not mask the other functions, just install the package but do not load it with library(); instead, just call the functions with the chroma:: prefix (e.g. chroma::rgb() instead of rgb(), to avoid masking rgb() in package grDevices).

Color parsing

Parse colors in various specifications

rgb(r=0.5, g=0.5, b=0.5)
ryb(r=0.5, y=0.5, b=0.5)
hsv(h=120, s=0.5, v=0.5)
hsl(h=120, s=0.5, l=0.5)
hsi(h=120, s=0.5, i=0.5)
hcl(h=120, c=0.5, l=0.5)
lab(l=0.5, a=-0.5, b=0.2)
hex("#393939")
hex("F39")
css("rgb(100,100,100)")
temperature(5600)
wavelength(500)

Parse directly from a matrix

x <- matrix(c(0.2, 0.5, 0.5, 0.5, 0.6, 0.4), ncol=3)
print(x)
rgb(x)

Color formating

Convert back in various formats

as.rgb("coral1")
as.ryb("coral1")
as.hsv("coral1")
as.hsl("coral1")
as.hsi("coral1")
as.hcl("coral1")
as.lab("coral1")
as.hex("coral1")
as.css("coral1")
as.temperature("coral1")
as.wavelength("coral1")

Color manipulation

Slightly modify a base color

col <- "#7BBBFE"
show_col(c(col, brighten(col), darken(col)))
show_col(c(col, desaturate(col), saturate(col)))

Make a color semi-transparent

col <- "#7BBBFE"
show_col(c(col, alpha(col), alpha(col, 0.2)))

Mix or blend two colors

show_col(c("#7BBBFE", "#FDFF68", mix("#7BBBFE", "#FDFF68")))
show_col(c("#7BBBFE", "#FDFF68", blend("#7BBBFE", "#FDFF68")))
show_col(c("#7BBBFE", "#FDFF68", blend("#7BBBFE", "#FDFF68", mode="screen")))

Compute the contrast between two colors

contrast("darkblue", "darkgreen")
contrast("yellow", "darkgreen")

Compute the Euclidean or perceptual distance between two colors (following the CIE Delta E 2000 formula)

# pick three colors; lightgreen is closer to darkgreen that darkblue is
show_col(c("darkblue", "darkgreen", "lightgreen"))
color_distance("darkblue", "darkgreen")
color_distance("lightgreen", "darkgreen")
CMClc("darkblue", "darkgreen")
CMClc("lightgreen", "darkgreen")
deltaE("darkblue", "darkgreen")
deltaE("lightgreen", "darkgreen")

which allows to find the closest perceived color in an array of possibilities

target <- "limegreen"
choices <- rainbow(10)
closest_color <- choices[which.min(deltaE(target, choices))]
show_col(target, choices, closest_color)

Extract or set a color channel

col <- "#7BBBFE"
channel(col, model="hcl", "h")
channel(col, model="hsv", "s")
channel(col, model="hsi", "i")
channel(col, model="rgb", "r")

col1 <- col2 <- col
channel(col1, model="hcl", "h") <- 120
channel(col2, model="hcl", "l") <- 1
show_col(c(col, col1, col2))

Compute or set the perceived luminance of a color

luminance(c("red", "yellow", "darkblue"))

col1 <- col2 <- col <- "#7BBBFE"
luminance(col)
luminance(col1) <- 0.6
luminance(col2) <- 0.2
show_col(c(col, col1, col2))

Color scales and palettes

All scales and palettes are organised the same way:

• functions ending in *_scale return a function that takes a numeric vector x as argument and returns the corresponding colors in the scale.
• functions ending in *_map are shortcuts that build the scale, map the values, and return the colors.
• functions ending in *_palette return a function that takes an integer n as argument and returns n equally spaced colors along the scale.
• functions ending in *_colors are shortcut that create the palette and return the n colors.

x <- 0:10/10
s <- interp_scale()
s(x)
# or
interp_map(x)

n <- 11
p <- interp_palette()
p(n)
# or
interp_colors(n)

Palettes can be built by interpolating between colors

show_col(interp_colors(10))
show_col(interp_colors(10, colors=c("#2D2B63", "#F7FF84")))
show_col(interp_colors(10, colors=c("#2D2B63", "#FB3C44", "#F7FF84")))

For the perception of "change" along the scale to be more linear, colors can be interpolated using bezier curves and their lightness can be corrected.

show_col(interp_colors(10, colors=c("#2D2B63", "#FB3C44", "#F7FF84")))
show_col(interp_colors(10, colors=c("#2D2B63", "#FB3C44", "#F7FF84"), interp="bezier"))
show_col(interp_colors(10, colors=c("#2D2B63", "#FB3C44", "#F7FF84"), interp="bezier", correct.lightness=TRUE))

Preset palettes are available, from colorbrewer

show_col(lapply(brewer_info$name, function(x) {brewer_colors(n=7, name=x)}))

Or viridis

show_col(
  viridis_colors(10),
  magma_colors(10),
  plasma_colors(10),
  inferno_colors(10)
)

Or cubehelix

show_col(
  cubehelix_colors(10),
  cubehelix_colors(10, h=300, rot=-0.75),
  cubehelix_colors(10, h=120, rot=0.5),
  cubehelix_colors(10, h=300, rot=0.5)
)

Or ETOPO1 and Wikipedia for topographical maps

show_col(
  etopo_colors(100),
  wikitopo_colors(100)
)

Custom, perceptually appropriate, palettes can be built in HCL space, for either discrete

show_col(hue_colors(10))

or continuous variables

show_col(chroma_colors(10, h=140))
show_col(light_colors(10, h=140))

Helper functions

Additional functions are included to facilitate the use of the color scales in base and ggplot2 plots.

sidemargin() and sidelegend() allow to place legends on the side of base plots

attach(iris)
# make larger right margin
pars <- sidemargin()
# plot the data with a nice color scale
plot(Petal.Length, Petal.Width, col=hue_map(Species), pch=19)
# add a legend for the scale on the side
sidelegend(legend=levels(Species), col=hue_colors(nlevels(Species)), pch=19)
# set graphical parameters back to default
par(pars)
detach(iris)

persp_facets() allows to compute the colors for each facet of a persp() plot

persp(maunga, theta=50, phi=25, scale=FALSE, expand=2,
      border=alpha("black", 0.4),
      col=magma_map(persp_facets(maunga$z)))

Various scale_color_*() and scale_fill_*() functions allow to use the color scales defined in chroma directly with ggplots and scale_xy_map() draws nice x and y scales for maps.

ggplot(iris) +
  geom_point(aes(Petal.Length, Sepal.Length, color=Species)) +
  scale_color_cubehelix_d(h=0, rot=0.75, c=1, l=c(0.6, 0.6))
ggplot(thaixyz) + coord_quickmap() +
 geom_raster(aes(x, y, fill=z)) +
 scale_fill_wikitopo() + scale_xy_map()

---

Happy coloring!

jiho/chroma documentation built on Nov. 26, 2022, 2:39 a.m.