Preparing plots for display

Overview

Teaching: 40 min
Exercises: 90 min

Questions

• How can I make my graphics effective?

Objectives

• To be able to use modify my plots to make them more understandable

• To change the looks of a plot.

• To combine multiple plots together in a figure

• To save plots to a file for use elsewhere.

Going back to our two roles a data visualisation might play:

• As a discovery tool, to explore a dataset
• As a storytelling tool, to highlight features of a dataset

The plots we have been producing so far are suitable for the first point, but are lacking the ability to be effective communication tools. To create visualisations that are ready to be displayed to others, whether in a publication or other form, we need to consider the non-data elements of our plots and make sure they contribute to our message.

We will continue working with our previous plot looking at the change in life expectancy over time for countries beginning with the letter “A”:

rough_plot <- ggplot(data = a_countries, aes(x = year, y = lifeExp, color = continent)) +
  geom_line() + 
  facet_wrap( ~ country)

Storing plots

You will see that we have stored the plot above into a variable named rough_plot. We can continue adding layers to this variable and they will be applied to the plot as usual. Saving a plot into a variable can be useful to save typing if you have a basic structure that you want to test adding different layers to.

Modifying text

To clean this figure up for a publication we need to make it a little more presentable. One of the first things we will do is clean up some of the labels for the different aesthetics. Currently, these are taken directly from the column names in the gapminder data frame; year, lifeExp, continent. This is fine when we are creating rough drafts to explore the dataset, but Year, Life Expectancy, and Continent would be more appropriate for a final figure.

We can do this by adding a new layer with the labs() function. Legend titles are set using the same names we used in the aes specification. Thus below the color legend title is set using color = "Continent", while the title of a fill legend would be set using fill = "MyTitle".

Within the labs() function we can also give the plot a title, subtitle, or caption to help explain what the plot is showing.

rough_plot +
  labs(
    title = "Figure 1",      # main title of figure
    x = "Year",              # x axis title
    y = "Life Expectancy",   # y axis title
    color = "Continent"      # title of legend
  ) 

Challenge 1

“Figure 1” is not a very useful title.

• Modify the code above so that the title is more descriptive
• Let’s also make sure we acknowledge our sources. Add a caption reading “Data source: Gapminder”

Solution

rough_plot +
  labs(
    title = "Growth in life expectancy for 'A' countries",
    caption = "Data source: Gapminder",
    x = "Year",              
    y = "Life Expectancy",  
    color = "Continent"      
  ) 

Modifying looks

We might also want to alter the theme of our plot to fit the visual style we are aiming for. The default theme that has been applied to our plots so far is theme_grey(), but there are a number of other complete themes inbuilt that you can use. For a more traditional look for publication, you may wish to try theme_bw(), theme_linedraw(), or theme_minimal().

rough_plot +
  labs(
    title = "Figure 1",
    x = "Year",              
    y = "Life expectancy",   
    color = "Continent"     
  ) +
  theme_bw()

We can modify individual elements of the theme of the plot using the theme() function. There are a lot of theme elements you can modify (see the function reference for a full list and examples of how to use them). For our plot, we may decide that we wish to make the fontface of the title bold and remove the minor gridlines from the axes.

  rough_plot +
    labs(
      title = "Figure 1",
      x = "Year",              
      y = "Life expectancy",   
      color = "Continent"     
    ) +
    theme_bw() +
    theme(
      panel.grid.minor = element_blank(),
      plot.title = element_text(face = "bold")
    )

Order matters

If you intend to modify an inbuilt theme, make sure to do it after adding the complete theme to the plot. Calling theme() and then adding a theme_bw() will override your changes with the theme_bw() settings.

In general, you remove a theme element by setting it to element_blank(). To change a theme element, you provide the appropriate settings in an element_text(), element_line(), element_rect(), etc. function. For example, to set the size of the axis lines, you would pass axis.line = element_line(size = 2) to the theme() function.

Challenge 2

What would you pass to the theme() function to make the following changes?

1) Remove the grey boxes behind the country names (controlled by the strip.background parameter).

2) Increase the size of the major gridlines (panel.grid.major) to 1 (need to use element_line())

3) Change the axis titles to be shrunk to size 10 and coloured blue

4) Position the legend at the bottom of the plot.

Solution

1) strip.background = element_blank()

2) panel.grid.major = element_line(size = 1)

3) axis.title = element_text(size = 10, colour = "blue")

4) legend.position = "bottom"

Exporting the plot

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 the plot from above, we first assign it to a variable lifeExp_plot, then tell ggsave to save that plot in png format to a directory called results. (Make sure you have a results/ folder in your working directory.)

lifeExp_plot <- rough_plot +
  labs(
    title = "Figure 1",
    x = "Year",              
    y = "Life expectancy",   
    color = "Continent"     
  ) +
  theme_bw() +
  theme(
    panel.grid.minor = element_blank(),
    plot.title = element_text(face = "bold")
  )

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

There are two nice things about ggsave. First, it defaults to the last plot, so if you omit the plot argument it will automatically save the last plot you created with ggplot. Secondly, it tries to determine the format you want to save your plot in from the file extension you provide for the filename (for example .png or .pdf). If you need to, you can specify the format explicitly in the device argument.

Multi-panel figures

Sometimes, one plot cannot tell the whole story and you need to combine multiple plots into a single figure. There are several ways you can perform this, but today we will be using the cowplot package

Installing cowplot

cowplot is not part of the tidyverse, and so you will need to install it separately if you want to make multi-panel figures. Install cowplot by running the following code once on your computer:

install.packages("cowplot")

And remember that you will need to load it with library() each time you wish to use it in a script:

library(cowplot)

The plot_grid() function from cowplot allows you to combine multiple ggplot objects into a single figure. Save each individual plot into a variable and provide them all as arguments to plot_grid():

library(cowplot)

# Making some plots
plot1 <- ggplot(gapminder, aes(x = gdpPercap, y = lifeExp)) + geom_point()
plot2 <- ggplot(gapminder, aes(x = continent, y = lifeExp)) + geom_boxplot()
plot3 <- ggplot(gapminder, aes(x = gdpPercap, y = pop)) + geom_point()
plot4 <- ggplot(gapminder, aes(x = lifeExp, y = pop)) + geom_point()

# Combining them into one
plot_grid(plot1, plot2, plot3, plot4)

By default, each panel is given the same size. You can adjust this using the rel_heights or rel_widths argument. Provide this argument a vector of numbers of relative sizes for the rows (rel_heights) or columns (rel_widths) in the final figure.

plot_grid(plot1, plot2, plot3, plot4, rel_heights = c(1, 3))

The subfigures can be labelled with the labels argument. You can either provide the exact labels or use "AUTO" for uppercase letters as labels, and "auto" for lowercase letters.

plot_grid(plot1, plot2, plot3, plot4, labels = "AUTO")

This is a taste of what you can do with ggplot2. RStudio provides a really useful cheat sheet of the different layers available, and more extensive documentation is available on the ggplot2 website. Finally, if you have no idea how to change something, a quick Google search will usually send you to a relevant question and answer on Stack Overflow with reusable code to modify!

Challenge 3

Tidy Tuesday is a weekly challenge to explore and visualise a dataset. One challenge was to work with data on board games that was used in an article at FiveThirtyEight

Working in small groups:

1) Map out the elements of the three graphs from the article

2) Download the data using

board_games <- readr::read_csv("https://raw.githubusercontent.com/rfordatascience/tidytuesday/master/data/2019/2019-03-12/board_games.csv")

3) Inspect the board_games data frame to determine which variables are required for each plot and if any further data manipulation is required

4) Reproduce the plots from the article. (If you want to get really serious, look into the ggthemes package for ggthemes::theme_fivethirtyeight())

5) Work individually to produce one new plot from the data. Show this figure to your group and discuss how effective it is at communicating a message.

Challenge: putting it all together

We have previously worked with meterological observations from the BOM. Let’s return to this data and explore it visually.

Getting the data

• Reopen your RStudio project working with the BOM meterological data OR
• Revisit the previous exercise for instructions on setting up and getting the data.

Question 1

For the Perth station (ID 9225), produce three scatter plots showing the relationship between the maximum temperature and each other measurement recorded (minimum temperature, rainfall and solar exposure).

Hint

Make sure your measurement values are all converted to numeric types with as.numeric(). They will have been imported as text values as part of the previous challenge.

To work with just the Perth measurements, make sure to filter() your dataset.

The only aesthetics you will need to map for this question are x and y. Then, a geom_point() layer will produce a scatterplot of the two mapped values.

Question 2

Display these four measurements for the Perth station in a single scatter plot by using additional aesthetic mappings.

You may need to try a few different data/aesthetic mappings to find one you like.

Hint

The aesthetics that geom_point() can use are found in the help page (?geom_point) under the Aesthetics heading

Hint 2

One possible solution is to map the temperature data to the x/y aesthetics, the rainfall to the size aesthetic, and the solar exposure to the colour aesthetic.

Question 3

Take the four plots you have produced in Q1 and Q2 and save them as a multi-panel figure.

Hint

The ggsave() function has a width and height argument to control how large a figure to create. This may be useful if your initial attempt at saving a figure looks too squashed.

Question 4

Using the entire BOM dataset, calculate the average monthly rainfall for each station. Produce a lineplot to visualise this data and the state each station is in.

Hint

For the data manipulation, you will need to use group_by() with multiple variables, and then summarise() to get the averaged rainfall measurements. Then join in a data frame containing the metadata about each station to be able to know which state each station is in.

If your lines in the plot are looking a bit odd, you may need to use the group aesthetic to make sure there is one line per station.

There are many ways you could use to integrate the state information – the colour and linetype aesthetics, or by using faceting. Which do you find makes for a better plot?

Question 5

Produce a plot of an interesting relationship you have found while exploring the dataset. Describe the message your plot is meant to convey to a viewer and the choices you made during the visualisation process to make your visualisation more effective at getting that message across.

Key Points

• Make sure your plots have useful titles, labels and legends

• Use theme() to modify how your plot looks

• The cowplot package lets you create a multi-plot figure

• Use ggsave() to save a plot to a file