Colors and color functions

SPDS, uni.kn

2023 01 01

unikn::

This vignette explains the colors, color palettes, and color-related functions provided by the unikn package. (See the vignettes on color recipes and institutional colors for more specialized tasks and the vignette on text for information on text boxes and decorations.)

Please install and/or load the unikn package to get started:

# install.packages('unikn')  # install unikn from CRAN client
library('unikn')             # loads the package

Overview

The unikn package provides some colors (e.g., Seeblau) and color palettes (e.g., pal_unikn). However, its functionality is mainly based on color-related functions that are useful beyond the colors and palettes of this package.

The package provides two main functions for interacting with color palettes: seecol() and usecol().

  1. seecol() is a general-purpose tool for seeing (inspecting or visualizing) colors or color palettes. The seecol() function takes two main arguments:

    • pal provides either one or multiple color palettes (with a default of pal = "unikn_all");
    • n specifies the number of desired colors (with a default of n = "all").

Based on the input of pal, the seecol() function distinguishes between two modes:

  1. usecol() allows using colors or color palettes (e.g., when creating visualizations) without showing its details. The usecol() function also takes arguments for conveniently manipulating color palettes:

    • pal provides either one or multiple color palettes (with a default of pal = pal_unikn);
    • n specifies the number of desired colors (with a default of n = "all");
    • alpha adjusts the opacity of all colors in pal (e.g., alpha = .50 for medium transparency).

Two additional functions allow finding colors by similarity or name:

  1. simcol() allows finding similar colors (given a target color, a set of candidate colors, and some tolerance threshold(s));

  2. grepal() allows finding colors with particular names (i.e., colors whose names match some pattern or regular expression);

Finally, some auxiliary functions support specific color functions:

The rest of this vignette provides examples of and some details on using these functions. (See the Color recipes vignette for more examples of solving color-related tasks.)

Viewing colors and color palettes with seecol()

The behavior of the seecol() function distinguishes between two modes and depends on the input to its initial pal argument. It either shows (A) the details of an individual color palette, or (B) allows comparing multiple color palettes. The next two sections will address both modes in turn.

A. Viewing details of a single color palette

When the pal argument of the seecol() function specifies a single color palette, the function plots a more detailed view of this particular color palette:

seecol(pal_unikn)  # view details of pal_unikn 

The detailed overview of a color palette provides us with

When a color palette contains too many colors, the HEX and RGB values are no longer printed. However, setting hex and rgb to TRUE will force them to be shown.

Note that seecol() also returns the color palette that is being shown. Thus, a typical workflow comprises both seeing a particular color palette and saving it (for storing and applying it later):

my_pal <- seecol(pal_unikn_light)  # view details of AND save a color palette  

Due to saving the color palette (here to my_pal) we can later use it in a visualization:

barplot(1/sqrt(1:10), col = my_pal)  # use my_pal in a plot

Note that seecol() invisibly returns the color palette.
Thus, the following will plot the palette pal_bordeaux without doing anything else with it:

seecol(pal_bordeaux)

but the following would both plot and assign the palette to my_pal:

my_pal <- seecol(pal_bordeaux)

B. Viewing and comparing multiple color palettes

The second mode of seecol() is invoked by providing (a list of) multiple color palettes to its pal argument.
In this case, the function allows comparing these palettes by plotting a color vector for each palette. Some special keywords within the unikn package denote sets of color palettes:

Calling seecol with pal set to these keywords allows comparing pre-defined sets of the color palettes:

Viewing the uni.kn color palettes:

seecol("unikn_all")  # all uni.kn color palettes

  1. three basic color palettes:
seecol("unikn_basic")

Note, that pal_unikn_web and pal_unikn_ppt are almost identical, but differ in how vibrant their colors are.

  1. three paired color palettes:
seecol("pair_all")

  1. all preferred colors from the spectrum and their respective gradients:
seecol("pref_all")

  1. only the pre-defined color gradients:
seecol("grad_all")

All additional color palettes (from other institutions):

seecol("add")

See the vignette on Institutional colors for creating color palettes for other institutions.

Other seecol() arguments

The seecol() function provides some aesthetic parameters for adjusting how color palettes are plotted:

Examples:

seecol("grad_all", col_brd = "black", lwd_brd = .5, main = "Color gradients (with black borders)")

seecol(pal_seegruen, col_brd = "white", lwd_brd = 5, main = "A color palette (with white borders)")

Using a color palette with usecol()

The usecol() function allows directly using a color palette in a plot (i.e., without first viewing it). usecol() corresponds to seecol() by taking the same main arguments (pal and n). However, as its purpose is using the colors specified by pal, rather than plotting (or seeing) them, its pal argument typically contains only one color palette:

# Using a color palette:
barplot(1/sqrt(1:11), col = usecol(pal_unikn))

Note that the seecol() and usecol() functions are both quite permissive with respect to specifying their pal argument: A particular color palette (e.g., pal_seegruen) can not only be displayed by providing it (as an object) but also by providing its name (i.e., "pal_seegruen") or even an incomplete object name or name (i.e., "seegruen" or seegruen). Hence, the following expressions all yield the same result:

seecol(pal_seegruen)
seecol("pal_seegruen")
seecol("seegruen")
seecol(seegruen)  # issues a warning, but works 

Customizing color palettes

Both the seecol() and the usecol() functions allow a flexible on-the-fly customization of color palettes.

Specifying a value for the n argument of seecol() an usecol() allows:

Passing a vector of colors and/or color palettes allows users to create and view their own color palettes.

Finally, specifying a value for alpha (in a range from 0 to 1) allows controlling the transparency of the color palette(s), with higher values for alpha corresponding to higher transparency (i.e., lower opacity).

Selecting subsets

Using only a subset of colors:

seecol("unikn_all", n = 4)

seecol(pal_unikn, 4)

Importantly, when using pre-defined color palettes of unikn but a value of n that is smaller than the length of the current color palette, usecol and seecol select a predefined subset of colors:

barplot(1/sqrt(1:2), col = usecol(pal_seeblau, n = 2))
barplot(1/sqrt(1:3), col = usecol(pal_seeblau, n = 3))

Extending color palettes

For values of n that are larger than the number of available colors in pal, the specified color palette is extended using ColorRampPalette:

seecol("unikn_all", n = 12)

Both seecol() and usecol() allow extending or truncating color palettes to a desired number n of colors. For instance:

seecol(pal_seeblau, n = 8)
barplot(1/sqrt(1:9), col = usecol(pal_bordeaux, n = 3))

Mixing colors and color palettes

By passing a vector to pal, we can concatenate 2 color palettes and connect them with a color (here: "white") as the midpoint of a new color palette:

seecol(pal = c(rev(pal_petrol), "white", pal_bordeaux))

We can combine a set of colors and extend this palette by specifying an n argument that is larger than the length of the specified palette:

seecol(pal = usecol(c(Karpfenblau, Seeblau, "gold"), n = 10))

These custom palettes can easily be used in a plot. For instance, we can define and use a subset of the pal_unikn_pair palette as follows:

my_pair <- seecol(pal_unikn_pair, n = 10)


# Create data: 
dat <- matrix(sample(5:10, size = 10, replace = TRUE), ncol = 5)

# Plot in my_pair colors:
barplot(dat, beside = TRUE, col = my_pair)

Creating linear color gradients is also supported by the shades_of() function (see below).

Controlling transparency

Both seecol() and usecol() accept an alpha argument (in a range from 0 to 1) for controlling the transparency of color palettes, with higher values for alpha corresponding to lower transparency (i.e., higher opacity).

Displaying a specific color palette at a medium opacity/transparency:

seecol(pal_unikn, alpha = 0.5)

Setting opacity for a custom color palette:

four_cols <- usecol(c("steelblue", "gold", "firebrick", "forestgreen"), alpha = 2/3)

seecol(four_cols, main = "Four named colors with added transparency")

Setting opacity for comparing of multiple color palettes:

seecol("grad", alpha = 0.67, main = "Seeing color palettes with added transparency")

Creating and comparing custom palettes

Suppose we want to compare a newly created color palette to existing color palettes. To achieve this, advanced users can use the seecol() function for displaying and comparing different custom palettes. When provided with a list of color palettes as the input to its pal argument, seecol() will show a comparison of the inputs:

# Define 2 palettes: 
pal1 <- c(rev(pal_seeblau), "white", pal_bordeaux)
pal2 <- usecol(c(Karpfenblau, Seeblau, "gold"), n = 10)

# Show the my_pair palette from above, the 2 palettes just defined, and 2 pre-defined palettes:  
seecol(list(my_pair, pal1, pal2, pal_unikn, pal_unikn_pair))

Note that unknown color palettes are named pal_\(n\), in increasing order. Palettes known to seecol() are labeled by their respective names. Labeling only custom palettes works by setting the pal_names argument to a character vector of appropriate length:

seecol(list(my_pair, pal1, pal2, pal_unikn, pal_unikn_pair), 
       pal_names = c("my_pair", "blue_bord", "blue_yell"),
       main = "Labeling custom color palettes")

If the pal_names argument is specified and corresponds to the length of all color palettes, the default names of all color palettes are overwritten by pal_names:

seecol(list(my_pair, pal1, pal2, pal_unikn, pal_unikn_pair), 
       pal_names = c("my_pair", "blue_bord", "blue_yell", "blue_black", "mix_pair"),
       main = "Comparing and labeling custom color palettes")

As before, we can use lower values of n for truncating/obtaining shorter subsets of color palettes:

seecol(list(my_pair, pal1, pal2, pal_unikn, pal_unikn_pair), n = 5)

or higher values of n for extending color palettes:

seecol(list(my_pair, pal1, pal2, pal_unikn, pal_unikn_pair), n = 15)

Finding colors

Two familiar color search tasks are addressed by the simcol() and grepal() functions:

Finding similar colors with simcol()

Assuming that our favorite color is "deeppink", a good question is: How can we find similar colors? Given some target color, the simcol() function searches through a set of colors to find and return visually similar ones:

simcol("deeppink", plot = FALSE)
#>     deeppink    deeppink2      maroon1      maroon2   violetred1   violetred2 
#>   "deeppink"  "deeppink2"    "maroon1"    "maroon2" "violetred1" "violetred2"

By default, simcol() searches though all named R colors of colors() (of the grDevices package), but adjusting the col_candidates and tol arguments allows for more fine-grained searches:

simcol("deepskyblue", col_candidates = pal_seeblau, tol = c(50, 50, 100))
#>   deepskyblue      seeblau3      seeblau4      seeblau5 
#> "deepskyblue"     "#59C7EB"     "#00A9E0"     "#008ECE"