3  Plotting and describing data

Before knowing how to plot something, you should know what you want to plot:

Scenario Which plot R command
Numeric Histogram or boxplot hist() andboxplot
Numeric with numeric Scatterplot plot
Numeric with categorical Boxplot boxplot(numeric~categorical)
Categorical with categorical mosaicplot or grouped barplot mosaicplot(table(categorical, categorical)) or barplot(data, beside=TRUE)

Here is a website with a decision tree about when to choose which plot https://www.data-to-viz.com/

Once you know what you want to plot, there are lot of websites that will show you the respective R code. One important consideration, however: there are at least two popular ways of doing R graphis:

for which the code looks completely different. You should probably get to know both types, but I would recommend to start with base R. I’m just saying this so that you are not confused, because a lot of the examples will also show you ggplot code. Here some useful links:

Thinking about how to do good figures

3.1 Plotting

3.1.1 One variable

3.1.1.1 Numerical variable - Histogram and Boxplot

  • The histogram plots the frequency of the values of a numerical variable with bins (otherwise each unique value will appear only once, the range will be cut in n elements). The number of bins is automatically inferred by the function but can be also changed by the user
  • The boxplot plots the distribution of a numerical variable based on summary statistics (the quantiles). The boxplot is particular useful for comparing/contrasting a numerical with a categorical variable (see below)
par(mfrow = c(1,2)) # number of plots, one row, two columns
hist(iris$Sepal.Length, 
     main = "Histogram", # title
     xlab = "Sepal.Length", 
     ylab = "Frequency",
     las = 1) # rotation of x and y values (las = 1, all of them should be horizontal)

boxplot(iris$Sepal.Length,      
        main = "Boxplot", # title
        ylab = "Values")

library(ggplot2)
library(ggpubr)

plt1 = 
  ggplot(iris, aes(x = Sepal.Length)) +
    geom_histogram() +
    ggtitle("Histogram") +
    xlab("Sepal.Length") +
    ylab("Frequency") +
    theme_bw() # scientific theme (white background)

plt2 = 
  ggplot(iris, aes(y = Sepal.Length)) +
    geom_boxplot() +
      ggtitle("Boxplot") +
      ylab("Values") +
      theme_bw() 


ggarrange(plt1, plt2, ncol = 2L, nrow = 1L)
## `stat_bin()` using `bins = 30`. Pick better value with `binwidth`.

3.1.1.2 Categorical variable - Barplot

str(mtcars)
## 'data.frame':    32 obs. of  11 variables:
##  $ mpg : num  21 21 22.8 21.4 18.7 18.1 14.3 24.4 22.8 19.2 ...
##  $ cyl : num  6 6 4 6 8 6 8 4 4 6 ...
##  $ disp: num  160 160 108 258 360 ...
##  $ hp  : num  110 110 93 110 175 105 245 62 95 123 ...
##  $ drat: num  3.9 3.9 3.85 3.08 3.15 2.76 3.21 3.69 3.92 3.92 ...
##  $ wt  : num  2.62 2.88 2.32 3.21 3.44 ...
##  $ qsec: num  16.5 17 18.6 19.4 17 ...
##  $ vs  : num  0 0 1 1 0 1 0 1 1 1 ...
##  $ am  : num  1 1 1 0 0 0 0 0 0 0 ...
##  $ gear: num  4 4 4 3 3 3 3 4 4 4 ...
##  $ carb: num  4 4 1 1 2 1 4 2 2 4 ...
counts = table(mtcars$cyl)
counts
## 
##  4  6  8 
## 11  7 14
barplot(counts, 
        main = "Barplot of Cyl",
        ylab = "Number of occurrences",
        xlab = "Cyl levels",
        col = "#4488AA")

ggplot(mtcars, aes(x = cyl)) +
  geom_bar(fill = "#4488AA") +
    ggtitle("Barplot of Cyl") +
    xlab("Number of occurrences") +
    ylab("Cyl levels") +
    theme_bw()

3.1.2 Two variables

The general idea of plotting is to look for correlations / associations between variables, i.e. is there a non-random pattern between the two variables.

3.1.2.1 Numerical vs numerical variable - Scatterplot

# Scatterplot
par(mfrow = c(1,2))
plot(airquality$Solar.R, airquality$Ozone)

# plot(Ozone ~ Solar.R, data = airquality) #the same

# different symbol for each month
plot(Ozone ~ Solar.R, data = airquality, pch = Month)

We can also add other objects such as lines to our existing plot:

par(mfrow = c(1,1))
plot(Ozone ~ Solar.R, data = airquality)
abline(h = 50)

plt1 = ggplot(airquality, aes(x = Solar.R, y = Ozone)) +
        geom_point() +
        theme_bw()

plt2 = ggplot(airquality, aes(x = Solar.R, y = Ozone)) +
        geom_point(shape = airquality$Month) +
        theme_bw()

ggarrange(plt1, plt2, ncol = 2L, nrow = 1L)
## Warning: Removed 42 rows containing missing values (`geom_point()`).
## Removed 42 rows containing missing values (`geom_point()`).

We can also add other objects such as lines to our existing plot:

ggplot(airquality, aes(x = Solar.R, y = Ozone)) +
  geom_point(shape = airquality$Month) +
  geom_abline(intercept = 50, slope = 0) +
    theme_bw()
## Warning: Removed 42 rows containing missing values (`geom_point()`).

3.1.2.2 Categorical vs numerical variable - Boxplot

Often we have a numerical variable (e.g. weight/fitness) and a categorical vairable that tells us the group of the observation (e.g. control or treatment). To compare visually now the distributions of the numerical variable between the levels of the grouping variable, we can use a boxplot

boxplot(mpg ~ cyl, mtcars, notch=TRUE) # formula notation
## Warning in (function (z, notch = FALSE, width = NULL, varwidth = FALSE, : some
## notches went outside hinges ('box'): maybe set notch=FALSE

# boxplot(x = mtcars$cyl, y = mtcars$mpg) # the same
ggplot(mtcars, aes(y = mpg, group = cyl)) +
  geom_boxplot(notch=TRUE) +
    theme_bw()
## Notch went outside hinges
## ℹ Do you want `notch = FALSE`?
## Notch went outside hinges
## ℹ Do you want `notch = FALSE`?