Customize diagrams: Formatting

TODO
Install required packages
Formatting elements for all kinds of diagrams
- Plot symbols and line types
- Diagram title, axis labels, axis limits, aspect ratio
Setting parameters with par()
Colors
Useful packages
Detach (automatically) loaded packages (if possible)
Get the article source from GitHub

TODO

link to diagAddElements -> axis(), diagMultiple

Install required packages

RColorBrewer

wants <- c("RColorBrewer")
has   <- wants %in% rownames(installed.packages())
if(any(!has)) install.packages(wants[!has])

Formatting elements for all kinds of diagrams

Nearly all R base diagrams come with a shared set of options to control typical diagram elements like the title, axis labels and limits, or the type of plot symbol. They are illustrated here using the plot() function for simple scatter plots.

Plot symbols and line types

Plot symbols are chosen with option pch.

X <- row(matrix(numeric(6*5), nrow=6, ncol=5))
Y <- col(matrix(numeric(6*5), nrow=6, ncol=5))

par(mar=c(1, 1, 4, 2))
plot(0:5, seq(1, 5, length.out=6), type="n", xlab=NA, ylab=NA,
     axes=FALSE, main="Data symbols")
points(X[1:26], Y[1:26], pch=0:25, bg="gray", cex=3)
text(X[1:26]-0.3, Y[1:26], labels=0:25)
text(0.2, 3, labels="symbols for pch", srt=90, cex=1.2)

Line types are chosen with option lty, line widths with option lwd, and round vs. square line ends with lend.

X <- row(matrix(numeric(6*12), nrow=6, ncol=12))
Y <- col(matrix(numeric(6*12), nrow=6, ncol=12))

par(mar=c(1, 1, 4, 2))
plot(0:6, seq(1, 12, length.out=7), type="n", axes=FALSE)
matlines(X[ , 1:6],  Y[ , 1:6],  lty=6:1, lwd=6:1, lend=0, col="blue")
matlines(X[ , 7:12], Y[ , 7:12], lty=6:1, lwd=6:1, lend=1, col="black")

## add annotations
text(rep(0.7, 12), Y[1, 1:12], labels=c(6:1, 6:1))
text(0, 7, labels="line types: lty, line widths: lwd", srt=90, cex=1.2)
text(0.32, 9, labels="line ending: lend=1", srt=90, cex=1.2)
text(0.32, 3, labels="line ending: lend=0", srt=90, cex=1.2)

Diagram title, axis labels, axis limits, aspect ratio

set.seed(123)
N <- 100
x <- rnorm(N, 100, 15)
y <- 0.3*x + rnorm(N, 0, 7)
plot(x, y, main="Scatter plot", sub="aspect ratio = 1",
     xlab="x axis", ylab="y axis",
     xlim=c(60, 140), asp=1, pch=16, col="blue")

Setting parameters with `par()`

Formatting details of a diagram can also be controlled with a call to par() before a diagram is opened. Some aspects can be set both with par(), and with the diagram function - e.g., with options to plot(). These include (see ?par for explanations):

cex, cex.axis, cex.main, cex.lab for the size of plot symbols, axis labels and diagram title
col for the color of plot symbols
font, family for the font of the diagram annotations
las for the orientation of axis labels
lend, lty, lwd, pch for the style of lines and plot symbols
xaxs, yaxs for the precise axis limits
xaxt, yaxt for the presence of axes

Other aspects can only be set in par(). These include: bt, mar, oma, xlog, ylog. The return value of par() is a list with the old value for the changed option. When it is saved, it can later be passed as an option to par() to reset the option for the current graphics device to its previous value.

par(mfrow=c(1, 2))
op <- par(col="gray60", family="serif", bty="n", mar=c(7, 5, 7, 1), pch=16)
plot(rnorm(10), main="Changed formatting")
par(op)
plot(rnorm(10), main="Standard format")

Colors

Default palette and other palette choices

N <- 6
(colDef <- palette()[1:N])

[1] "black"   "red"     "green3"  "blue"    "cyan"    "magenta"

(colAll <- sample(colors(), N, replace=FALSE))

[1] "darkviolet"    "ivory"         "snow3"         "mediumorchid3"
[5] "grey97"        "lightsalmon"

colRain <- rainbow(N)
colTopo <- topo.colors(N)
colCm   <- cm.colors(N)
colHeat <- heat.colors(N)

len    <- 1/N
xLeft  <- rep(seq(0,   1-len, by=len), times=N)
yBot   <- rep(seq(0,   1-len, by=len),  each=N)
xRight <- rep(seq(len, 1,     by=len), times=N)
yTop   <- rep(seq(len, 1,     by=len),  each=N)

par(mar=c(0, 4, 1, 0) + 0.1)
plot(c(0, 1), c(0, 1), axes=FALSE, xlab=NA, ylab=NA, type="n",
     asp=1, main="Colors")
rect(xLeft, yBot, xRight, yTop, border=NA,
     col=c(colDef, colAll, colRain, colTopo, colCm, colHeat))
par(xpd=NA)
text(-0.05, seq(0, 1-len, length.out=N) + len/2, adj=1,
     labels=c("default", "sample", "rainbow", "topo", "cm", "heat"))

RGB colors and transparency

rgb(0, 1, 1)

[1] "#00FFFF"

rgb(t(col2rgb("red")/255))

[1] "#FF0000"

rgb(1, 0, 0, 0.5)

[1] "#FF000080"

N  <- 150
xx <- rnorm(N, 100, 15)
yy <- 0.4*xx + rnorm(N, 0, 10)
plot(xx, yy, pch=16, cex=3.5, col=rgb(0, 0, 1, 0.3))
points(xx-20, yy-20, pch=16, cex=3.5, col=rgb(1, 0, 0, 0.3))
points(xx+20, yy-20, pch=16, cex=3.5, col=rgb(0, 1, 0, 0.3))

Other color spaces

hsv(0.1666, 1, 1)

[1] "#FFFF00"

rgb2hsv(matrix(c(0, 1, 1), nrow=3))

         [,1]
h 0.500000000
s 1.000000000
v 0.003921569

hcl(h=120, c=35, l=85)

[1] "#BBDEB1"

gray(0.5)

[1] "#808080"

Color palettes with `RColorBrewer`

N <- 6
library(RColorBrewer)
(bPal <- brewer.pal(N, "Blues"))

[1] "#EFF3FF" "#C6DBEF" "#9ECAE1" "#6BAED6" "#3182BD" "#08519C"

colorRampPalette(bPal)(15)

 [1] "#EFF3FF" "#E0EAF9" "#D1E1F3" "#C3D9EE" "#B4D3E9" "#A6CDE4" "#96C6DF"
 [8] "#84BCDB" "#72B2D7" "#5EA4D0" "#4994C7" "#3585BE" "#2574B3" "#1662A7"
[15] "#08519C"

b1 <- colorRampPalette(brewer.pal(N, "Reds"))(N)
b2 <- colorRampPalette(brewer.pal(N, "Greens"))(N)
b3 <- colorRampPalette(brewer.pal(N, "BuGn"))(N)
b4 <- colorRampPalette(brewer.pal(N, "BrBG"))(N)
b5 <- colorRampPalette(brewer.pal(N, "OrRd"))(N)
b6 <- colorRampPalette(brewer.pal(N, "Accent"))(N)

len    <- 1/N
xLeft  <- rep(seq(0,   1-len, by=len), times=N)
yBot   <- rep(seq(0,   1-len, by=len),  each=N)
xRight <- rep(seq(len, 1,     by=len), times=N)
yTop   <- rep(seq(len, 1,     by=len),  each=N)

par(mar=c(0, 4, 1, 0) + 0.1)
plot(c(0, 1), c(0, 1), axes=FALSE, xlab=NA, ylab=NA, type="n",
     asp=1, main="Colors")
rect(xLeft, yBot, xRight, yTop, border=NA,
     col=c(b1, b2, b3, b4, b5, b6))

Useful packages

Package colorspace provides more functions for converting between different color spaces. Its documentation illustrates many color palettes for different use cases.
The R color chart gives a very nice overview of colors available in R.
Package imager can read bitmap images and has many functions for image analysis / manipulation

Detach (automatically) loaded packages (if possible)

try(detach(package:RColorBrewer))

Get the article source from GitHub

R markdown - markdown - R code - all posts