Data visualization with ggplot2

Slides in full screen     Download PDF slides

1 Task

Note

Find the solution here.

1.1 Get started

A helpful resource to consult for this task can be the ggplot2 cheatsheet.

Remember to put library(tidyverse) (or library(ggplot2)) on top of your script to access the ggplot functions.

To practice plotting with the ggplot2 package, we will use a data set on 3 species of penguins on islands in Antarctica.

Artwork by Allison Horst

The data is available from the palmerpenguins package. To get it, you have to:

  • Install the package with install.packages("palmerpenguins")
  • Load the package at the beginning of your script with library(palmerpenguins)

The data set is called penguins:

  • The data set is available after you loaded the package
  • Just type penguins in the console and you should see the penguins tibble
  • Find a description of the variables in the help page ?penguins

Take a moment to get familiar with the data set and its variables.

1.2 Exploratory plots

Explore the data set and it’s variables with ggplot. Below, you find some suggestions for plots. You can start with the plot type your are most interested in and then continue from there. You don’t have to finish all the plots If you have your own ideas for interesting plots with the penguin data set, feel free to deviate from the tasks.

In this exploratory section, don’t worry about the beauty of your plots. This task is about exploring the data and testing different visualization options.

1.2.1 Relationship between bill length and bill depth (scatterplot)

What is the relationship between bill length and bill depth?

  • Create a scatterplot with bill length on the x-axis and bill depth on the y-axis
  • Can you add a regression line?
  • Add species as color aesthetic. Does your interpretation of the data change?
    • What is the difference between adding color as a global aesthetic or as a local aesthetic of the point layer?
  • Add species as shape aesthetic to distinguish the species

1.2.2 Difference in flipper length between species (boxplot)

Is there a difference in flipper length between the species?

  • Create a boxplot of the flipper length (y-axis) for the different species (x-axis)
  • Try adding notches to the boxplots
  • Extra: Add a layer with geom_point(). Try setting position = position_jitter() as argument in the point layer. What does it do?

1.2.3 Differences between body mass of male and female penguins (boxplot)

Are male penguins heavier than female penguins? And is this different between the 3 species?

  • Create a boxplot with body mass on the y axis and sex on the x axis
  • Add the difference between penguin species to this plot. Try the different options ggplot offers
    • Species as color aesthetic
    • Species as fill aesthetic
    • Species as facet using facet_wrap
  • Extra: What happens if you use geom_violin instead of geom_boxplot? Can you combine both geoms in one plot?

1.2.4 Distribution of flipper length between species (histogram)

Make a histogram of the the flipper length separated by species.

Try different methods of separating the species (color or facet).

Compare stacked and overlapping histograms.

1.2.5 Penguin flipper length by species and sex (heatmap)

Create a heat map that shows:

  • The categories sex and species on x- and y-axis
  • The flipper length as color

1.3 Beautify the plots

First, choose one of the tasks 1.3.1 or 1.3.2, then do task 1.4 on saving plots. If you still have time, you can come back to the task you didn’t do.

But also here, if you have other ideas, feel free to deviate from the tasks.

1.3.1 Beautify the plots from Task 1

Take a plot you did in the previous task and make it look nicer.

Here’s a list of ideas:

  • Add a theme layer
  • Customize the theme, e.g.
    • Change the position of the legend
    • Make the axis titles bold
  • Change the color/fill scale of the plot
    • Use scale_color_manual or scale_fill_manual
    • Try scale_color_viridis_d() or scale_fill_viridis_d() with different options
    • Try a color scale from the paletteer package
  • Change the labels of the x- and y-axis and add a title to the plot
  • Make the points transparent, give them a different shape, …

1.3.2 Can you reproduce this plot?

Take a look at this plot:

Visualization from the palmerpenguins package

It is similar to the one from Task 1.3.1 but more beautiful. Can you reproduce this plot?

The colors that are used are:

penguin_cols <- c("darkorange", "purple", "cyan4")

1.4 Save one of the plots on your machine

Save one of the plots you produced in a variable and then use ggsave to save the plot on your machine. Save the plot in a dedicated plot directory in your RStudio project.

Note: Make sure the path where you save the image exists. If you e.g. want to save in img/, then you first have to create the img folder in your project directory. For this you can use the Files pane of RStudio.

1.5 References

Horst AM, Hill AP, Gorman KB (2020). palmerpenguins: Palmer Archipelago (Antarctica) penguin data. R package version 0.1.0. https://allisonhorst.github.io/palmerpenguins/. doi: 10.5281/zenodo.3960218.