Processing math: 100%
+ - 0:00:00
Notes for current slide
Notes for next slide

Base R graphics

R Foundations for Data Analysis

Marcin Kierczak, Nima Rafati

1/24

Example graphics

Life expectancy.

2/24

Exampel graphics

3/24

Graphical devices

The concept of a graphical device is crucial for understanding R graphics.

4/24

Graphical devices

The concept of a graphical device is crucial for understanding R graphics.
A device can be a screen (default) or a file.

4/24

Graphical devices

The concept of a graphical device is crucial for understanding R graphics.
A device can be a screen (default) or a file.

Creating a plot entails:

  • opening a graphical device (not necessary for plotting on screen),
4/24

Graphical devices

The concept of a graphical device is crucial for understanding R graphics.
A device can be a screen (default) or a file.

Creating a plot entails:

  • opening a graphical device (not necessary for plotting on screen),
  • plotting to the graphical device,
4/24

Graphical devices

The concept of a graphical device is crucial for understanding R graphics.
A device can be a screen (default) or a file.

Creating a plot entails:

  • opening a graphical device (not necessary for plotting on screen),
  • plotting to the graphical device,
  • closing the graphical device (very important!)
4/24

Graphical devices

The concept of a graphical device is crucial for understanding R graphics.
A device can be a screen (default) or a file.

Creating a plot entails:

  • opening a graphical device (not necessary for plotting on screen),
  • plotting to the graphical device,
  • closing the graphical device (very important!)

Common graphical devices

Device.Type Functions Notes
Screen default device Displays plot on screen
Bitmap/Raster png(), bmp(), tiff(), jpeg() Use for images with pixel data
Vector svg(), pdf() High-quality, scalable graphics
Cairo cairo_pdf(), cairo_ps() Enhanced quality, especially on Windows

For more information visit this link.

4/24

Working with graphical devices

png(filename = 'myplot.png', width = 320, height = 320, antialias = T)
# anti-aliasing, a technique that smooths the edges of shapes and text in the plot, making them appear less pixelated or jagged.
plot(x=c(1,2,7), y=c(2,3,5))
dev.off()
5/24

Working with graphical devices

png(filename = 'myplot.png', width = 320, height = 320, antialias = T)
# anti-aliasing, a technique that smooths the edges of shapes and text in the plot, making them appear less pixelated or jagged.
plot(x=c(1,2,7), y=c(2,3,5))
dev.off()

What we did was, in sequence:

  • open a graphical device, here png() with some parameters,
  • do the actual plotting to the device using plot() and
  • close the graphical device using dev.off().
5/24

Working with graphical devices

png(filename = 'myplot.png', width = 320, height = 320, antialias = T)
# anti-aliasing, a technique that smooths the edges of shapes and text in the plot, making them appear less pixelated or jagged.
plot(x=c(1,2,7), y=c(2,3,5))
dev.off()

What we did was, in sequence:

  • open a graphical device, here png() with some parameters,
  • do the actual plotting to the device using plot() and
  • close the graphical device using dev.off().

It is of paramount importance to remember to close graphical devices. Otherwise, our plots may end up in some random weird files or on screens.

5/24

Standard graphical device viewport

source: rgraphics.limnology.wisc.edu

6/24

Standard graphical device viewport

source: rgraphics.limnology.wisc.edu

c(bottom, left, top, righ)

6/24

base::plot() basics

base::plot() is a basic command that lets you visualize your data and results. It is very powerful yet takes some effort to fully master it. Let's begin with plotting three points:

  • A(1,2);
  • B(2,3);
  • C(7,5);
plot(x=c(1,2,7), y=c(2,3,5))

7/24

Anatomy of a plot — episode 1

For convenience, we will create a data frame with our points:

##   x y
## A 1 2
## B 2 3
## C 7 5
plot(df)

8/24

Anatomy of a plot — episode 2

There is many parameters one can set in plot(). Let's have a closer look at some of them:

  • pch – type of the plotting symbol
  • col – color of the points
  • cex – scale for points
  • main – main title of the plot
  • sub – subtitle of the plot
  • xlab – X-axis label
  • ylab – Y-axis label
  • las – axis labels orientation
  • cex.axis – axis lables scale
  • xlim – X-axis limits or range
  • ylim – Y-axis limits or range

Let's make our plot a bit fancier...

# recycling rule in action!
plot(df, pch = c(15, 17, 19), col = c("tomato", "slateblue"), main = "Three points",
    sub = "Stage 1")

9/24

Graphical parameters

Graphical parameters can be set in two different ways:

  • as plotting function arguments, e.g. plot(dat, cex = 0.5)
  • using par() to set parameters globally
# read current graphical parameters
par()
# first, save the current parameters so that you
# can set them back if needed
old_par <- par() # should work in theory, practise varies :-(
# now, modify what you want
par(cex.axis = 0.8, bg='grey')
# do your plotting
plot(................)
# restore the old parameters if you want
par(old_par)
10/24

The pch parameter

11/24

How to make the pch cheatsheet

# create a grid of coordinates
coords <- expand.grid(1:6,1:6)
# make a vector of numerical pch symbols
pch.num <- c(0:25)
# and a vector of character pch symbols
pch.symb <- c('*', '.', 'o', 'O', '0', '-', '+', '|', '%', '#')
# plot numerical pch
plot(coords[1:26,1], coords[1:26,2], pch=pch.num,
bty='n', xaxt='n', yaxt='n', bg='red',
xlab='', ylab='')
# and character pch's
points(coords[27:36,1], coords[27:36,2], pch=pch.symb)
# label them
text(coords[,1], coords[,2], c(1:26, pch.symb), pos = 1,
col='slateblue', cex=.8)

Now, make your own cheatsheet for the lty parameter!

12/24

Layers

Elements are added to a plot in the same order you plot them. It is like layers in a graphical program. Think about it! For instance the auxiliary grid lines should be plotted before the actual data points.

# make an empty plot
plot(1:5, type = "n", las = 1, bty = "n")
grid(col = "grey", lty = 3)
points(1:5, pch = 19, cex = 3)
abline(h = 3, col = "red")

The line overlaps one data point. It is better to plot it before plotting points()
13/24

Some thoughts about plotting

There is a few points you should have in mind when working with plots:

14/24

Some thoughts about plotting

There is a few points you should have in mind when working with plots:

  • raster or vector graphics,
14/24

Some thoughts about plotting

There is a few points you should have in mind when working with plots:

  • raster or vector graphics,
  • colors, e.g. color-blind people, warm vs. cool colors and optical illusions (check colorbrewer2.org),
14/24

Some thoughts about plotting

There is a few points you should have in mind when working with plots:

  • raster or vector graphics,
  • colors, e.g. color-blind people, warm vs. cool colors and optical illusions (check colorbrewer2.org),
  • avoid complications, e.g. 3D plots, pie charts etc.,
14/24

Some thoughts about plotting

There is a few points you should have in mind when working with plots:

  • raster or vector graphics,
  • colors, e.g. color-blind people, warm vs. cool colors and optical illusions (check colorbrewer2.org),
  • avoid complications, e.g. 3D plots, pie charts etc.,
  • use black and shades of grey for things you do not need to emphasize, i.e. basically everything except your main result,
14/24

Some thoughts about plotting

There is a few points you should have in mind when working with plots:

  • raster or vector graphics,
  • colors, e.g. color-blind people, warm vs. cool colors and optical illusions (check colorbrewer2.org),
  • avoid complications, e.g. 3D plots, pie charts etc.,
  • use black and shades of grey for things you do not need to emphasize, i.e. basically everything except your main result,
  • avoid 3 axes on one figure.
14/24

Many plots on one figure

par(mfrow=c(2,3))
plot(1:5)
plot(1:5, pch=19, col='red')
plot(1:5, pch=15, col='blue')
hist(rnorm(100, mean = 0, sd=1))
hist(rnorm(100, mean = 7, sd=3))
hist(rnorm(100, mean = 1, sd=0.5))
par(mfrow=c(1,1))

Alternative: use par(mfcol=c(3,2)).

15/24

Many plots on one figure

par(mfrow=c(2,3))
plot(1:5)
plot(1:5, pch=19, col='red')
plot(1:5, pch=15, col='blue')
hist(rnorm(100, mean = 0, sd=1))
hist(rnorm(100, mean = 7, sd=3))
hist(rnorm(100, mean = 1, sd=0.5))
par(mfrow=c(1,1))

Alternative: use par(mfcol=c(3,2)).

15/24

Using graphics::layout()

M <- matrix(c(1,2,2,2,3,2,2,2), nrow = 2, ncol = 4, byrow = T)
M
layout(mat = M)
layout.show(n = 3) #It shows the layout setup without plots
##      [,1] [,2] [,3] [,4]
## [1,]    1    2    2    2
## [2,]    3    2    2    2
16/24

Using graphics::layout()

M <- matrix(c(1,2,2,2,3,2,2,2), nrow = 2, ncol = 4, byrow = T)
M
layout(mat = M)
layout.show(n = 3) #It shows the layout setup without plots
##      [,1] [,2] [,3] [,4]
## [1,]    1    2    2    2
## [2,]    3    2    2    2
M <- matrix(c(1,2,2,2,3,2,2,2), nrow = 2, ncol = 4, byrow = T)
layout(mat = M)
plot(1:5, pch=15, col='blue')
hist(rnorm(100, mean = 0, sd=1))
plot(1:5, pch=15, col='red')

16/24

Defining colors

mycol <- c(rgb(0, 0, 1), "olivedrab", "#FF0000")
plot(1:3, c(1, 1, 1), col = mycol, pch = 19, cex = 3)
points(2, 1, col = rgb(0, 0, 1, 0.5), pch = 15, cex = 5)
17/24

Defining colors

mycol <- c(rgb(0, 0, 1), "olivedrab", "#FF0000")
plot(1:3, c(1, 1, 1), col = mycol, pch = 19, cex = 3)
points(2, 1, col = rgb(0, 0, 1, 0.5), pch = 15, cex = 5)

17/24

Color palettes

There some built-in palettes: default, hsv, gray, rainbow, terrain.colors, topo.colors, cm.colors, heat.colors

mypal <- heat.colors(10)
mypal
pie(x = rep(1, times = 10), col = mypal)
##  [1] "#FF0000" "#FF2400" "#FF4900" "#FF6D00" "#FF9200" "#FFB600" "#FFDB00"
##  [8] "#FFFF00" "#FFFF40" "#FFFFBF"

18/24

Custom color palettes

You can easily create custom palettes:

mypal <- colorRampPalette(c("red", "green", "blue"))
pie(x = rep(1, times = 12), col = mypal(12))
class(mypal)
## [1] "function"

19/24

Custom color palettes

You can easily create custom palettes:

mypal <- colorRampPalette(c("red", "green", "blue"))
pie(x = rep(1, times = 12), col = mypal(12))
class(mypal)
## [1] "function"

Note:grDevices::colorRampPalette() returns a function for generating colors based on the defined custom palette!

There is an excellent package RColorBrewer that offers a number of pre-made palettes, e.g. color-blind safe palette.

Package wesanderson offers palettes based on Wes Anderson's movies :-)

19/24

Box-and-whiskers plot — theory

There are also more specialized R functions used for creating specific types of plots. One commonly used is graphics::boxplot()

Rule of thumb. If median of one boxplot is outside the box of another, the median difference is likely to be significant α=5%.

20/24

Box-and-whiskers plot

boxplot(decrease ~ treatment,
data = OrchardSprays,
log = "y", col = "bisque",
varwidth=TRUE)

21/24

Box-and-whiskers plot

boxplot(decrease ~ treatment,
data = OrchardSprays,
log = "y", col = "bisque",
varwidth=TRUE)

The add=TRUE parameter in plots allows you to compose plots using previously plotted stuff!

attach(InsectSprays)
boxplot(count~spray, data=InsectSprays[spray %in% c("C","F"),], col="lightgray")
boxplot(count~spray, data=InsectSprays[spray %in% c("A","B"),], col="tomato", add=TRUE)
boxplot(count~spray, data=InsectSprays[spray %in% c("D","E"),], col="slateblue", add=TRUE)
detach(InsectSprays)

21/24

Plotting categorical data

data(Titanic) # Load the data
vcd::mosaic(Titanic,
            labeling = vcd::labeling_border(rot_labels = c(0,0,0,0)))

Package vcd provides a lot of ways of visualizing categorical data.

22/24

Plotting simple heatmaps

heatmap(matrix(rnorm(100, mean = 0, sd = 1), nrow = 10),
col=terrain.colors(10))

23/24

Graph gallery

For more awesome examples visit: http://www.r-graph-gallery.com

24/24

Thank you! Questions?

Graphics from
Created: 31-Oct-2024 • SciLifeLab • NBIS

24/24

Example graphics

Life expectancy.

2/24
Paused

Help

Keyboard shortcuts

↑, ←, Pg Up, k Go to previous slide
↓, →, Pg Dn, Space, j Go to next slide
Home Go to first slide
End Go to last slide
Number + Return Go to specific slide
b / m / f Toggle blackout / mirrored / fullscreen mode
c Clone slideshow
p Toggle presenter mode
t Restart the presentation timer
?, h Toggle this help
Esc Back to slideshow