In R, you can choose different ways to plot your data and results. By default, R comes with built-in graphical functions. Other functions, or entirely different paradigms, can be enabled using different packages. Here we give a brief overview of the standard plotting functions that are useful for quick data exploration. These are also referred to as R “base graphics” as no extra package is required.
The function plot(x, y, ...) is the most basic way to plot points defined by their x- and y-coordinates. Here, x and y are vectors of equal length containing the x- and y-coordinates of the points to be plotted. For example,
# generate data
x <- 1:30 # creates a vector of 30 values that we use as x-coordinates
y <- rnorm(30, mean = x) # creates a vector of 30 values that we use as y-coordinates; each values is drawn from normal distribution with with an increasing mean value
plot(y) # if no x-vector is given the default x = (1, 2, 3, ...) is used
plot(x, y) # this plot produces an identical result but strictly speaking x is not necessary here
# there are many different option one can add to the plot commandWithin this basic plotting command, we can add many additional arguments to change to the exact layout of the plot. For a more thorough overview of the options see here.
plot(y, type = "l") # using lines instead of points
plot(y, type = "l", lty = 2) # change the line-type, the value 2 produces a dashed line.
plot(y, type = "b") # using both lines and points
plot(y, type = "b", pch = 4) # change the type of point
plot(y, type = "b", pch = 4, col = "blue") # change color
plot(y, col = "orange", xlab = "my x-label", ylab = "a y-label") # add axes labels
plot(y, xlim = c(1,35), ylim = c(1,40)) # set plot ranges
plot(y, log="y") # plots the same data but with the y-axis on a log-scale (the same can be done with the x-axis)In the above examples, we plotted the entries of a vector v without specifying values for the x-axes of the plot. Indeed, if no x-vector is specified, then the default x = (1, 2, 3, …) is used. This does not work anymore if y-values do not correspond to x-values that are consecutive natural numbers. Here an example:
x <- seq(0, 5, 0.5)
x
y <- x^2
plot(x, y) # check the x-axes; this is the desired plot
plot(y) # without specifying x, the y-values are at wrong positionsR base graphics are drawn interactively: you can start with a plot, and then overlay other elements on the existing plot.
x <- 1:30
y <- rnorm(30, mean = x)
plot(x, y, type = "l", ylim = c(1,60), col = "orange", ylab = "f(x)")
lines(x, 2*y, type = "l", col = "blue") # adds a line to the previous plot
legend("topleft", legend = c("f(x)=y","f(x)=2y"), lty = c(1,1), col = c("orange", "blue")) # adds a plot legend
abline(h = 20) # adding a horizontal line by setting a y-value h
abline(v = 10) # adding a vertical line by setting an x-value v
abline(a = 0, b = 1, col = "red") # add a line to the same plot by specifying its intercept a and slope bWe can plot the graphs of functions using the command curve(expr, from, to, ...), where expr is a mathematical expression containing an argument x or function name, and from and to defines the range of x-values for which the function is plotted. This is a useful tool to visually explore the shape of functions.
curve(exp, from = -3, to = 3) # plot of the exponential function in the interval (-3, 3)
abline(v = 0) # adding a vertical line at x = 0
logistic <- function(n, r, k){# definition of the user-defined function giving the population growth rate of a logistically growing population as a function the current population size n, the intrinsic growth rate r and the carrying capacity k
n*r*(1-n/k)
}
curve(logistic(x , r = 1, k = 2), from = 0, to =2.5) # plot of the function logistic as a function of the population size n = x
# note that the function curve requires that you name the argument that you use for plotting x
abline(h = 0) # adding a horizontal line at y = 0The plotting command matplot(x, y, ...) is an easy way to superimpose several data series. Here, x is a vector containing the values for the x-axes and y is a matrix containing the data as a function of x. Note, that matplot(x, y, ...) plots each column of y as separate data series.
x <- 1:10 # creates a vector of length 10, this will become the x-axes
y.1 <- x^2 # another vectors of length 10
y.2 <- x^2 - 2*x # and a third vector of length 10
matplot(x, cbind(y.1, y.2), col =c(2, 4), pch = c(0, 2), xlab = "x", ylab = "y.1 = x^2 and y.2 = x^2-2x")
legend(1, 100, legend = c("y.1","y.2"), col =c(2, 4), pch = c(0, 2)) # adds a plot legendIn the following example, the variable x is given by the first column of a matrix while the actual data are taken from the second and third column of that same matrix.
x <- 1:10 # creates a vector of length 10, this will become the x-axes
y.1 <- x^2 # two more vectors of length 10
y.2 <- x^2 - 2*x
data_matrix <- cbind(x, y.1, y.2) # combine the above three vectors into a matrix
data_matrix
matplot(data_matrix[ , 1], data_matrix[ , 2-3], pch = c(16, 1), xlab = "x", ylab = "y.1 = x^2 and y.2 = x^2-2*x")
legend(1, 100, legend = c("y.1","y.2"), pch = c(16, 1), col = c(1, 2)) The function hist(x, ...) is used to produce simple histograms, where x is a vector of values for which the histogram is desired. Histograms show the abundance or frequency of data falling in certain discrete classes.
# generate data, drawn from Poisson distribution with mean 3
d1 <- rpois(100, lambda = 3)
# basic histogram
hist(d1) # without further options R decides how to bin the data
# specify desired number of bins
hist(d1, breaks = 4) # four break points results in five bins
# specify bin edges
hist(d1, breaks = c(0, 1, 3, 5, 7, 11, 21)) # manually defining bins
# use frequencies (default if bins have equal size)
hist(d1, freq = TRUE)
# use density (= relative frequencies)
hist(d1, freq = FALSE)
# get the histogram, but without plotting it
z <- hist(d1, plot = FALSE)
# access elements of the histogram
z$counts # counts per bin
z$mids # midpoints of the binsA bar plot is used to represent data for discrete groups. Here are some examples using the function barplot(height, ...):
data(rivers) # load data on length of North American rivers
barplot(rivers, xlab = "rivers", ylab = "length")
barplot(rivers, xlab = "rivers", ylab = "length", horiz = TRUE) # horizontal
data(iris) # load iris data set
structure(iris) # inspect the iris data set
# color by group level
barplot(height = iris$Petal.Width, beside = TRUE, col = iris$Species, xlab = "species", ylab = "petal width")Box plots are used to show the range of a distribution, and the location of the bulk of its mass. Use the function boxplot(x, ...) where x specifies the data from which the boxplot is to be produced.
data(iris)
# set specific colors
boxplot(iris$Petal.Width ~ iris$Species, col = c("red", "green", "blue"))