PRESENTATION OVERVIEW

In this presentation...

• Introduce the three primary plotting paradigms in R for producing publication-quality plots and graphics

R has three MAIN plotting paradigms

• Base R graphics - The focus of this presentation

• The Lattice graphics package

  • Implementation of Trellis graphics suite of data visualization tools

  • Focused on multivariate data

• The ggplot2 graphics package

  • Plotting system based on the grammar of graphics

  • Tries to take the good parts of base R and lattice graphics and none of the bad parts

  • Takes care of many details that make plotting a hassle

  • Provides powerful tools to make it easy to produce complex, multi-layered graphics

Other packages extend R for specific types of plots

• d3heatmap

• rgl

• ggvis

• plot3d

• networkd3

• leaflet

• DiagrammeR

• timeseries

More helpful plotting resources

• 15 Questions All R Users Have About Plots

  • This article includes several of same tips discussed in this presentation
  • But, also shows how to accomplish many of the same effects using the ggplot2 or lattice graphics packages

• R Graphics Cookbook

  • Similar to the article above, only much more in-depth

CREATING PLOTS IN BASE R GRAPHICS WITH plot(x, y,...)

When the plot(x, y,...) function is executed

• The following events occur

  • A new plot window is created - more on this later

  • A new plot device is activated inside the plot window

  • The values defined in the vector x are stored as the \(x\) coordinates of the points \((x_i,y_i), i=1,...,n\)

  • The values defined in the vector y are stored as the \(y\) coordinates of the points \((x_i,y_i), i=1,...,n\)

  • If x and y contain the same number of elements the points \((x_i,y_i), i=1,...,n\) are printed in the active plot device

• Additional arguments, denoted by ... are then executed to change the active plot device settings, these may include

  • xlim = c(lower, upper) \(\rightarrow\) upper and lower limits of x-axis

  • ylim = c(lower, upper) \(\rightarrow\) upper and lower limits of y-axis

  • xlab = "text" \(\rightarrow\) title of x-axis

  • ylab = "text" \(\rightarrow\) title of y-axis

  • main = "text" \(\rightarrow\) title of plot

  • axes = TRUE \(\rightarrow\) show or suppress plot axes

  • yaxt = 's' or 'n' \(\rightarrow\) show or suppress y-axis values

  • xaxt = 's' or 'n' \(\rightarrow\) show or suppress x-axis values

  • ann = TRUE \(\rightarrow\) show or suppress ALL plot annotations

  • type = 'p' \(\rightarrow\) Change the type of objects being plotted

  • pch = 1 \(\rightarrow\) Set the type of plot character

  • lwd = 1 \(\rightarrow\) Set the line width

  • lty = 1 \(\rightarrow\) Change the line type

  • cex = 1 \(\rightarrow\) Set the character expansion value

  • col = 1 \(\rightarrow\) Change the color of plotted objects

• If the above arguments aren't specified, R will choose values for them based on the objects being plotted

Plotting Shiny App

• The shiny app below shows how the arguments to plot(x, y,...) effect the appearance of a plot

• For this app both x and y are vectors of length \(10\) with values randomly sampled from the set of integers \(1 - 50\)

  • x <- sample(1:50, size = 10)

  • y <- sample(1:50, size = 10)

• It's unlikely that x = y since the number of distinct samples of size \(n=10\) that could have been generated is

\[ \begin{aligned} \binom{n}{k} &= \frac{n!}{k! \, (n - k)!}\\\\ \binom{50}{10} &= \frac{50!}{10! \, 40!} =1.0272278\times 10^{10} \;\text{possible samples} \end{aligned} \]

Adding Objects To An Active Plot Device

After the plot(x, y,...) function runs

• The plot device remains active so additional objects can be added

  • Add text with text(x,...)

  • Add equations with expression(...)

  • Add points with points(x,...)

  • Add lines with lines(x,...) or abline(a,b, h, v, coef, reg,...)

  • Add a legend with legend( )

  • The device closes when code is run that doesn't affect the plot

Changing The Global Plot Settings With par(...)

Recall what happens when plot(x, y,...) is executed...

• A new plot window is created

• A new plot device is activated inside the plot window

• The values defined in the vector x are stored as the \(x\) coordinates of the points \((x_i,y_i), i=1,...,n\)

• The values defined in the vector y are stored as the \(y\) coordinates of the points \((x_i,y_i), i=1,...,n\)

• If x and y contain the same number of elements the points \((x_i,y_i), i=1,...,n\) are printed in the active plot device

• Additional arguments, denoted by ... are then executed to change the active plot device settings, these may include

There are also global plot settings

• That can only be changed with the par(...) function

  • xlog = FALSE

  • ylog = FALSE

  • adj = 0.5

  • ann = TRUE

  • ask = FALSE

  • bg = "white"

  • bty = "o"

  • cex = 1

  • cex.axis = 1

  • cex.lab = 1

  • cex.main = 1.2

  • cex.sub = 1

  • cin = c(0.15, 0.20)

  • col = "black"

  • col.axis = "black"

  • col.lab = "black"

  • col.main = "black"

  • col.sub = "black"

  • cra = c(14.4, 19.2)

  • crt = 0

  • csi = 0.2

  • cxy = c(0.03702180, 0.08792819)

  • din = c(5.291667, 4.114583)

  • err = 0

  • family = ""

  • fg = "black"

  • fig = c(0, 1, 0, 1)

  • fin = c(5.291667, 4.114583)

  • font = 1

  • font.axis = 1

  • font.lab = 1

  • font.main = 2

  • font.sub = 1

  • lab = c(5, 5, 7)

  • las = 0

  • lend = "round"

  • lheight = 1

  • ljoin = "round"

  • lmitre = 10

  • lty = "solid"

  • lwd = 1

  • mai = c(1.02, 0.82, 0.82, 0.42)

  • mar = c(5.1, 4.1, 4.1, 2.1)

  • mex = 1

  • mfcol = c(1, 1)

  • mfg = c(1, 1, 1, 1)

  • mfrow = c(1, 1)

  • mgp = c(3, 1, 0)

  • mkh = 0.001

  • new = FALSE

  • oma = c(0, 0, 0, 0)

  • omd = c(0, 1, 0, 1)

  • omi = c(0, 0, 0, 0)

  • page = TRUE

  • pch = 1

  • pin = c(4.051667, 2.274583)

  • plt = c(0.1549606, 0.9206299, 0.2478987, 0.8007089)

  • ps = 12

  • pty = "m"

  • smo = 1

  • srt = 0

  • tck = NA

  • tcl = -0.5

  • usr = c(0, 1, 0, 1)

  • xaxp = c(0, 1, 5)

  • xaxs = "r"

  • xaxt = "s"

  • xpd = FALSE

  • yaxp = c(0, 1, 5)

  • yaxs = "r"

  • yaxt = "s"

  • ylbias = 0.2

Plotting Complex Objects And Special Plot Functions

plot(x, y,...) can be used to display much more than just sets of points

• In the previous outline of what happens when plot(x, y,...) executes, some steps were purposefully skipped

• plot(x, y,...) is a "generic" function known as an S3 method in R parlance

• When plot(objects) is called, R first checks the class( ) and typeof( ) object being plotted

  • If "objects" means the vectors x and y, plot(x, y,...) calls plot.xy( )

  • If "objects" means a matrix, plot(x, y,...) calls plot.matrix( )

  • R does this automatically...you only need to use plot(object)

• In the packages that install with base R, there are several specific plot functions

  • plot.stepfun( ) example: plot(stepfun(1:3, c(1,2,4,3)))

  • plot.ecdf( ) example: plot(ecdf(rnorm(24)))

  • plot.table( ) example: plot(Titanic)

Several "wrapper" functions use plot(x, y,...) to create special purpose plots

- `plot.function( )` example: `curve(sin, -2*pi, 2*pi, xname = "t")`