2.1 RGB — What Computers See

Every color on screen is an additive mixture of Red, Green, and Blue light. Values run 0–255 (or 0–1) per channel. RGB is perfect for rendering but terrible for human reasoning: #1A6FAD and #5B9DC9 look nothing like “the same blue at different brightness” even though they are.

2.2 HSV — Artist’s Intuition

Hue–Saturation–Value rearranges RGB into something more intuitive:

HSV is better than RGB for picking colors by hand but is not perceptually uniform — equal steps in HSV do not look equal to the human eye.

2.3 HCL — Perceptually Uniform and Human-Friendly

Hue–Chroma–Luminance is the gold standard for data visualization:

Equal steps in HCL look equal to human eyes. This means palettes built in HCL space automatically have balanced visual weight across colors.

2.4 LUV (CIELUV) — The Math Behind HCL

HCL is the polar form of the CIELUV color space. LUV uses Cartesian coordinates (L, U, V) where L is luminance and the (U, V) plane encodes chromaticity. We will plot all of R’s 657 named colors in this space in Section 4.

3.1 Named Colors

# R has 657 named colors
length(colors())

## [1] 657

head(colors(), 20)

##  [1] "white"         "aliceblue"     "antiquewhite"  "antiquewhite1"
##  [5] "antiquewhite2" "antiquewhite3" "antiquewhite4" "aquamarine"   
##  [9] "aquamarine1"   "aquamarine2"   "aquamarine3"   "aquamarine4"  
## [13] "azure"         "azure1"        "azure2"        "azure3"       
## [17] "azure4"        "beige"         "bisque"        "bisque1"

3.2 Hex Codes and Conversion

# Hex → RGB
col2rgb("#268bd2")

##       [,1]
## red     38
## green  139
## blue   210

# RGB → Hex
rgb(38, 139, 210, maxColorValue = 255)

## [1] "#268BD2"

# Named color → Hex
col2rgb("steelblue")

##       [,1]
## red     70
## green  130
## blue   180

# HSV → Hex
hsv(h = 0.61, s = 0.73, v = 0.82)

## [1] "#386CD1"

# HCL → Hex (base R)
hcl(h = 220, c = 80, l = 55)

## [1] "#0097C1"

3.3 Plotting Named Colors by Hue

R’s 657 named colors sorted by Hue–Saturation–Value:

set_text_contrast <- function(color) {
  avg <- mean(col2rgb(color))
  ifelse(avg > 127, "black", "white")
}

txt_contrast <- sapply(colors(), set_text_contrast)

rgb_mat   <- col2rgb(colors())
hsv_mat   <- rgb2hsv(rgb_mat[1,], rgb_mat[2,], rgb_mat[3,], maxColorValue = 255)
hue_order <- order(hsv_mat[1,], hsv_mat[2,], hsv_mat[3,])

col_count <- 15
row_count <- ceiling(length(colors()) / col_count)

plot(0, type = "n", axes = FALSE, xlab = "", ylab = "",
     xlim = c(0.5, col_count + 0.5), ylim = c(row_count + 0.5, 0.5))
title("R Named Colors — Sorted by Hue, Saturation, Value", cex.main = 0.9)

for (j in seq(0, row_count - 1)) {
  for (i in seq(1, col_count)) {
    k <- j * col_count + i
    if (k <= length(colors())) {
      idx <- hue_order[k]
      rect(i - 0.5, j + 0.5, i + 0.5, j - 0.5,
           col = colors()[idx], border = NA)
      text(i, j, idx, cex = 0.35, col = txt_contrast[idx])
    }
  }
}

The index number shown on each tile is its position in colors() — use colors()[n] to retrieve any color by index.

Using colorspace in ggplot2

library(dplyr)

# qualitative: scale_color_discrete_qualitative()
# sequential:  scale_fill_continuous_sequential()
# diverging:   scale_fill_continuous_diverging()

set.seed(1)
demo_df <- data.frame(
  x     = rnorm(120),
  y     = rnorm(120),
  group = rep(LETTERS[1:6], 20),
  value = rnorm(120)
)

ggplot(demo_df, aes(x = x, y = y, color = group)) +
  geom_point(size = 2.5, alpha = 0.8) +
  scale_color_discrete_qualitative(palette = "Dark 3") +
  labs(title = "qualitative_hcl via scale_color_discrete_qualitative()",
       color = "Group") +
  theme_gray(base_size = 13)

You can inspect any palette’s HCL trajectory with specplot():

specplot(sequential_hcl(9, "Blues 3"), asp = 1/3)

The three lines show Hue (H), Chroma (C), and Luminance (L) as the palette progresses. A good sequential palette has monotonically increasing L and decreasing C — this is what makes it readable in grayscale.

Key ggplot2 scales

# Qualitative fill: scale_fill_brewer(palette = "Set2")
# Sequential fill:  scale_fill_distiller(palette = "Blues")
# Diverging fill:   scale_fill_distiller(palette = "RdBu")

ggplot(demo_df, aes(x = x, y = y, color = group)) +
  geom_point(size = 2.5, alpha = 0.85) +
  scale_color_brewer(palette = "Set2") +
  labs(title = "ColorBrewer Set2 — scale_color_brewer(palette = 'Set2')",
       color = "Group") +
  theme_gray(base_size = 13)

Choosing a ColorBrewer palette

# Which palettes are colorblind-safe?
subset(brewer.pal.info, colorblind == TRUE)

##         maxcolors category colorblind
## BrBG           11      div       TRUE
## PiYG           11      div       TRUE
## PRGn           11      div       TRUE
## PuOr           11      div       TRUE
## RdBu           11      div       TRUE
## RdYlBu         11      div       TRUE
## Dark2           8     qual       TRUE
## Paired         12     qual       TRUE
## Set2            8     qual       TRUE
## Blues           9      seq       TRUE
## BuGn            9      seq       TRUE
## BuPu            9      seq       TRUE
## GnBu            9      seq       TRUE
## Greens          9      seq       TRUE
## Greys           9      seq       TRUE
## Oranges         9      seq       TRUE
## OrRd            9      seq       TRUE
## PuBu            9      seq       TRUE
## PuBuGn          9      seq       TRUE
## PuRd            9      seq       TRUE
## Purples         9      seq       TRUE
## RdPu            9      seq       TRUE
## Reds            9      seq       TRUE
## YlGn            9      seq       TRUE
## YlGnBu          9      seq       TRUE
## YlOrBr          9      seq       TRUE
## YlOrRd          9      seq       TRUE

7.1 The HCL Hue Strip

library(tidyr)

h_vals <- seq(0, 360, length.out = 721)
fills  <- hcl(h_vals, c = 70, l = 55)
fills[is.na(fills)] <- "#AAAAAA"

strip_df <- data.frame(h = h_vals, fill = fills)

# Anchor points for the three palettes defined in Section 8
harm_df <- data.frame(
  h       = c(220, 40,                   # Tide: complementary
               30,  150, 270,             # Grove: triadic
               20,  110, 200, 290),       # Studio: tetradic
  palette = c(rep("Tide (Comp. 180°)", 2),
              rep("Grove (Tri. 120°)",  3),
              rep("Studio (Tet. 90°)",  4))
)

ggplot() +
  geom_tile(data = strip_df,
            aes(x = h, y = 0, fill = fill), width = 0.5, height = 0.55) +
  geom_point(data = harm_df,
             aes(x = h, y = 0, shape = palette),
             size = 4.5, color = "white", stroke = 1.8) +
  scale_fill_identity() +
  scale_x_continuous(breaks = seq(0, 360, 30),
                     labels = paste0(seq(0, 360, 30), "°"),
                     expand = c(0, 0)) +
  scale_shape_manual(values = c(21, 22, 24)) +
  labs(title = "HCL Hue Strip — Anchor Points for Three Project Palettes",
       x = "Hue (degrees)", shape = NULL) +
  theme_minimal(base_size = 12) +
  theme(
    axis.title.y    = element_blank(),
    axis.text.y     = element_blank(),
    panel.grid      = element_blank(),
    legend.position = "bottom"
  )

7.2 Complementary — 2 Hues, 180° Apart

Maximum contrast. Ideal for binary comparisons (treated vs control, before vs after). Risk of visual vibration when used at full saturation — reduce chroma or luminance on one member.

7.3 Triadic — 3 Hues, 120° Apart

Balanced and vibrant. Good for three-category data. All three hues share the same HCL chroma and luminance so no color dominates.

7.4 Tetradic (Square) — 4 Hues, 90° Apart

Rich four-color palettes. Harder to balance — typically keep chroma moderate (50–65) to avoid visual noise. One color should dominate and the others support.

7.5 Analogous — Adjacent Hues (30–60° Apart)

Low contrast, high harmony. Good for backgrounds, shading, or when you need many steps of the same “temperature.” Not suitable as a qualitative palette because the colors are too similar.

Quick-reference table

9.1 Qualitative — scale_color_manual() / scale_fill_manual()

# Three-group example using Grove (triadic)
set.seed(7)
df3 <- data.frame(
  x     = rnorm(90),
  y     = rnorm(90),
  group = rep(c("Alpha", "Beta", "Gamma"), 30)
)

ggplot(df3, aes(x = x, y = y, color = group)) +
  geom_point(size = 3, alpha = 0.85) +
  scale_color_manual(values = unname(grove)) +
  labs(title = "Grove palette — scale_color_manual(values = unname(grove))",
       color = "Group") +
  theme_gray(base_size = 13)

9.2 Two-Group Contrast — Tide Complementary Pair

df2 <- data.frame(
  group = rep(c("Control", "Treated"), each = 40),
  value = c(rnorm(40, 10, 2), rnorm(40, 13, 2.5))
)

ggplot(df2, aes(x = group, y = value, fill = group)) +
  geom_boxplot(alpha = 0.85, width = 0.5) +
  geom_jitter(aes(color = group), width = 0.12, size = 1.8, alpha = 0.6) +
  scale_fill_manual(values  = c(Control = tide[["tide_mid"]],
                                Treated = tide[["tide_amber"]])) +
  scale_color_manual(values = c(Control = tide[["tide_deep"]],
                                Treated = tide[["tide_gold"]])) +
  labs(title = "Tide palette — complementary pair for binary comparison") +
  guides(fill = "none", color = "none") +
  theme_gray(base_size = 13)

9.3 Four-Category — Studio Tetradic

df4 <- data.frame(
  category = rep(c("Q1", "Q2", "Q3", "Q4"), each = 30),
  x        = rnorm(120),
  y        = c(rnorm(30, 1), rnorm(30, -1), rnorm(30, 1, 0.5), rnorm(30, -0.5))
)

ggplot(df4, aes(x = x, y = y, color = category)) +
  geom_point(size = 2.5, alpha = 0.8) +
  scale_color_manual(values = unname(studio)) +
  labs(title = "Studio palette — scale_color_manual(values = unname(studio))",
       color = "Quarter") +
  theme_gray(base_size = 13)

9.4 Storing Palettes for Reuse

Define your palettes once in a project-level colors.R file and source() it at the top of each script:

# colors.R — project palette definitions

tide <- c(
  tide_deep  = "#1B4E8A",   # adjust to your actual hex values
  tide_mid   = "#4A82B8",
  tide_amber = "#C96A1A",
  tide_gold  = "#E8AB68"
)

grove <- c(
  grove_copper = "#A84C25",
  grove_fern   = "#2D7A52",
  grove_violet = "#4D3B8C"
)

studio <- c(
  studio_coral = "#B54230",
  studio_olive = "#5E7A2A",
  studio_steel = "#1E5E88",
  studio_plum  = "#5A3576"
)