ggplot

1 Creating graphs

Artwork by @allison_horst

1.1 Making graphs with ggplot2

Set up

We will use the palmer penguin data to make graphs.

If you don’t have a project open already, open your saved Rproj in Rstudio by clicking File menu then Open project and look for your .Rproj file. Click the Files tab in the bottom right pane, click the data folder and check the penguins.csv is in there. Then read it in.

penguins <- read.csv(file = "data/penguins.csv")

There are different plotting systems in R for example, the base plotting system. But the function ggplot in the ggplot2 package is used by the majority of people.

To use it you will need to ensure you have the ggplot2 package installed. Hint: you could use the Packages tab, bottom right and then load it using the function library().

library(ggplot2)

ggplot2 is built on the idea that any plot can be made from the same set of components: a data set, a mapping, and some geoms.

The key to understanding ggplot2 is to think about a graph in layers.

Let’s start off with an example:

ggplot(data = penguins, mapping = aes(x = body_mass_g, y = flipper_length_mm)) +
  geom_point()

In the command we call the ggplot function. Any of the arguments we give the ggplot function in the brackets following, will apply to all layers on the plot.

We’ve passed in two arguments to ggplot. First, we tell ggplot what data we want to use to make the figure. In this example, that is the penguins data we read in earlier.

For the second argument called mapping, we passed in the aes function, which tells ggplot how variables in the data map to aesthetic properties of the figure. In this case the x and y locations. Here we told ggplot we want to plot the “body mass” column of the penguins data frame on the x-axis, and the “flipper length” column on the y-axis.

By itself, the call to ggplot isn’t enough to draw a figure:

ggplot(data = penguins, mapping = aes(x = body_mass_g, y = flipper_length_mm))

We need to tell ggplot how we want to visually represent the data, which we do by adding a geom layer. In our example, we used geom_point, which tells ggplot we want to visually represent the relationship between x and y as a scatterplot of points:

ggplot(data = penguins, mapping = aes(x = body_mass_g, y = flipper_length_mm)) +
  geom_point()

Challenge 1

Modify the example so that the figure shows how bill length relates to flipper length:

ggplot(data = penguins, mapping = aes(x = body_mass_g, y = flipper_length_mm)) + 
    geom_point()

Solution to Challenge 1

Here is one possible solution:

ggplot(data = penguins, mapping = aes(x = bill_length_mm, y = flipper_length_mm)) + 
  geom_point()

Note: we can add arguments to geom_point to control the size, transparency and shape of the points.

ggplot(data = penguins, mapping = aes(x = bill_length_mm, y = flipper_length_mm)) + 
  geom_point(size = 5, alpha = 0.5, shape = "diamond")

Challenge 2

So far we’ve used the aes function to tell the scatterplot geom about the x and y locations of each point. Another aesthetic property we can modify is color. Add a colour= argument to the aes brackets to color the points according to the “species” column.

Solution to Challenge 2

Here is one possible solution:

ggplot(data = penguins, mapping = aes(x = bill_length_mm, y = flipper_length_mm, colour = species)) + 
  geom_point()

We can make other types of graph using other geoms. If we wanted to compare body mass among the penguin species we might use a boxplot. We can change the geom to do this:

ggplot(data = penguins, mapping = aes(x = species, y = body_mass_g)) + 
  geom_boxplot()

Tip: ggplot cheat sheet

There is a ggplot cheat sheet to download which you can look at to see what the geoms are for common types of graph, as well as other ggplot code.

Note that the layers on the cheat sheet have single letters and a + to represent the rest of the code for the graph.

1.2 Layers

If we want boxplots with data points on top to visualise the spread of the data, we can add another layer to the plot. geom_jitter() is good for this because it does not plot the points in one vertical line but randomly spreads them out:

ggplot(data = penguins, mapping = aes(x = species, y = body_mass_g)) + 
  geom_boxplot() +
  geom_jitter()

More layers can be added to control the way a graph looks. For example, we can change the axes labels:

ggplot(data = penguins, mapping = aes(x = species, y = body_mass_g)) + 
  geom_boxplot() +
  geom_jitter() +
  labs(x = "Species",
       y = "Body Mass (g)")

We could use lots of layers to control the background, gridlines, axis lines. However, using a theme instead does this in one layer. theme_classic() gives styling that might be found in a scientific journal.

ggplot(data = penguins, mapping = aes(x = species, y = body_mass_g)) + 
  geom_boxplot() +
  geom_jitter() +
  labs(x = "Species",
       y = "Body Mass (g)") +
  theme_classic()

Challenge 3

Create a ggplot graph using the geom_violin() which shows the difference in bill length among the sexes. Add layers for axes labels and theme styling.

Solution to Challenge 3

Here is one possible solution:

ggplot(data = penguins, mapping = aes(x = sex, y = bill_length_mm)) + 
  geom_violin() +
  labs(x = "Sex",
       y = "Bill Length (mm)") +
  theme_classic()

Note: when we learn about data wrangling we will go through ways of removing data points such as the penguins that had NA under sex.

1.3 Colour

We previously used colour in aes to give each species a different colour:

ggplot(data = penguins, mapping = aes(x = bill_length_mm, y = flipper_length_mm, colour = species)) + 
  geom_point()

Putting the colour = species in the aes function in the geom layer also works:

ggplot(data = penguins, mapping = aes(x = bill_length_mm, y = flipper_length_mm)) + 
  geom_point(aes(colour = species))

However, if we want to specify that all points should be a blue colour, oppose to levels in a variable being different colours, we use colour = but outside of the aes brackets:

ggplot(data = penguins, mapping = aes(x = bill_length_mm, y = flipper_length_mm)) + 
  geom_point(colour = "blue")

Note the word blue is put in “speech marks”.

Challenge 4

Choose a colour from this website that lists R colours and create a violin plot outlined in that colour which shows the difference in body mass among species. You can include layers for axes labels and styling.

Solution to Challenge 4

Here is one possible solution:

ggplot(data = penguins, mapping = aes(x = species, y = body_mass_g)) + 
  geom_violin(colour = "darkorchid")

Note: adding fill = "darkorange" will add a colour to the inside of the violin.

ggplot(data = penguins, mapping = aes(x = species, y = body_mass_g)) + 
  geom_violin(colour = "darkorchid", fill = "darkorange")

This graph uses default colours:

ggplot(data = penguins, mapping = aes(x = bill_length_mm, y = flipper_length_mm, colour = species)) + 
  geom_point() +
  labs(x = "Species",
       y = "Body Mass (g)") +
  theme_classic()

But to specify colours, use scale_colour_manual:

ggplot(data = penguins, mapping = aes(x = bill_length_mm, y = flipper_length_mm, colour = species)) + 
  geom_point() +
  labs(x = "Species",
       y = "Body Mass (g)") +
  theme_classic() +
  scale_colour_manual(values = c("darkorange","purple","cyan4"))

Tip: creating plots

When creating graphs, people adapt code from the internet by replacing the names of the data and variables in the example code with their data.

If searching for ggplot layers that will control specific parts of your graph, using the correct graph terminology helps. For example, tick marks, levels, error bars, axes, legends.

1.4 Plotting statistics

ggplot2 makes it easy to overlay statistical models represented by a trend line on a graph by adding a geom_smooth layer:

ggplot(data = penguins, mapping = aes(x = bill_length_mm, y = bill_depth_mm)) + 
  geom_point() +
  geom_smooth(method="lm")

Note: the lm stands for linear model and is the function used to run a linear regression.

We can also do this after the data has been split by species using group:

ggplot(data = penguins, mapping = aes(x = bill_length_mm, y = bill_depth_mm, group = species)) + 
  geom_point() +
  geom_smooth(method="lm")

Challenge 4

Make a scatterplot of bill length and bill depth, grouped and coloured by species, with trendlines for each species and triangles for points and specify colours (you can choose colours). Give it appropriate axes labels and a classic theme.

Solution to Challenge 4

Here is one possible solution:

ggplot(data = penguins, mapping = aes(x = bill_length_mm, y = bill_depth_mm, colour = species, grouped = species)) + 
  geom_point(shape = "triangle") +
  labs(x = "Bill Length (mm)",
       y = "Bill Depth (mm)") +
  theme_classic() +
  scale_colour_manual(values = c("darkorange","purple","cyan4")) +
  geom_smooth(method = lm)

## `geom_smooth()` using formula 'y ~ x'

1.5 Exporting plots

The ggsave() function allows you to export a plot created with ggplot. You can specify the dimension and resolution of your plot by adjusting the appropriate arguments (width, height and dpi) to create high quality graphics for publication.
In order to save a plot from above, we first assign it to a variable, we could call it penguin_plot:

penguin_plot <- ggplot(data = penguins, mapping = aes(x = bill_length_mm, y = bill_depth_mm, colour = species, grouped = species)) + 
  geom_point(shape = "triangle") +
  labs(x = "Bill Length (mm)",
       y = "Bill Depth (mm)") +
  theme_classic() +
  scale_colour_manual(values = c("darkorange","purple","cyan4")) +
  geom_smooth(method = lm)

Now we can tell ggsave to save that plot in our results folder, with the file name “billScatterplot”, in png format and specify the width and height in cm and the resolution (dpi). Check you have a results/ folder in your working directory by clicking the files tab in the bottom left pane.

ggsave(filename = "results/billScatterplot.png", plot = penguin_plot, width = 12, height = 10, dpi = 300, units = "cm")

Source

Adapted from R for Reproducible Scientific Analysis licensed CC_BY 4.0 by The Carpentries