Intro to paletteNMR

This package is intended to be used for storing and accessing NMR color palettes in R, for easy integration in visualization and reporting.

At the moment, it contains three core functions. The first, nmr_colors, produces a named list of colors (or, optionally) a contrasting set of colors. Use ?nmr_colors to see further documentation. The second, draw_palette, is a thin wrapper on nmr_colors for drawing ggplot figures using the palettes.

Then there are additional helper functions, including scale_* functions that integrate NMR palettes with ggplot2

Navigating existing palettes

First, a high-level view of the palettes available in the NMR set of palettes. At the top layer, there are two scales, categorical and numerical. Each scale contains various palettes, detailed below. And most palettes contain multiple sets, e.g. varying the base color.

See below for details on what sets to consider in specific applications (e.g. high contrast, B&W printing, etc.)

Categorical

For maximum legibility, you should have five or fewer categories. If possible, consider combining categories. Scales include:

Pairs
Quads
Map (for points and polygons)
Lines
hex and Hept

Quads

Among the quads, any consecutive subset of colors in these sets could be used since they maintain high contrast as indicated by the B&W variants. Quads C, F and H contain distinct colors, allowing them to be used on graphs of different attributes. Quad C or Hept E are best in case of B&W printing.

Lines

The line palette maximizes legibility when printed in B&W, but avoid having more than 4 categories in one line chart. Consider using unique line types or point markers if possible.

Maps

To improve legibility, you should use different shapes as well as colors for different category markers. Using a color to show water is optional–usually not needed (and best omitted)

Numerical

Non-divergent

Do not use the green and red sets side-by-side. If you need a lighter scale, select a longer sequence than you need and omit the darkest color. As with categorical palettes, avoid having more than five categories.

Avoid interpolating colors unless you need a continuous scale. If you do need continuous scale, use the viridis scale built into the ggplot2 package (see ?scale_color_viridis_d)

You can join a non-divergent series to itself with light colors on the end to accentuate a mean (i.e. the inverse of the divergent scale below)

Divergent

If you need a continuous scale, consider using the BrBG ramp from ColorBrewer, also in the ggplot2 package (see ?scale_color_brewer)

Accessing existing palettes

Calling nmr_colors returns the color IDs and their hex codes.

output <- nmr_colors(scale = "categorical", pal = "quad", set = "F")

output

##        B2        T1        G2        T3 
## "#115B6B" "#FFFFCC" "#649429" "#6B5E0A"

Create graphics using above call to nmr_colors with the following code:

data.frame(hex = output) |> 
  ggplot(aes(fill = hex)) +
  facet_wrap( ~ hex, ncol = 1) +
  geom_rect(aes(xmin = 1, xmax = 2,
                ymin = 1, ymax = 2)) +
  lims(x=c(1,2),y=c(1,2)) +
  scale_fill_manual(values = as.character(output),
                    guide = "none") +
  theme(axis.text = element_blank(),
        axis.ticks = element_blank(),
        panel.grid = element_blank(),
        panel.spacing = unit(-0.5, "lines"),
        plot.margin = unit(c(0,-0.25,0,-0.25), "lines"),
        strip.background = element_blank(),
        strip.text = element_blank())

paletteNMR also contains scale_* functions modeled on those in the ggplot2 package for visualization.

data.frame(hex = output) |> 
  ggplot(aes(fill = hex)) +
  facet_wrap( ~ hex, ncol = 1) +
  geom_rect(aes(xmin = 1, xmax = 2,
                ymin = 1, ymax = 2)) +
  lims(x=c(1,2),y=c(1,2)) +
  scale_nmr_fill_categorical(pal = "quad",
                             set = "F",
                             guide = "none") +
  theme(axis.text = element_blank(),
        axis.ticks = element_blank(),
        panel.grid = element_blank(),
        panel.spacing = unit(-0.5, "lines"),
        plot.margin = unit(c(0,-0.25,0,-0.25), "lines"),
        strip.background = element_blank(),
        strip.text = element_blank())

Drawing palettes (draw_palette)

draw_palette(scale = "numerical", "nondiverg", "RT5")

See the “palette_browse.Rmd” vignette to browse available palettes.