A primer of the ggplot2 package

Before you start


Learning objectives

Learn ggplot2 basics to create simple figures.


Table of contents

  1. Understanding the data structure of sf

Tips to make the most of the lecture notes


Interactive navigation tools

  • Click on the three horizontally stacked lines at the bottom left corner of the slide, then you will see table of contents, and you can jump to the section you want

  • Hit letter “o” on your keyboard and you will have a panel view of all the slides


Running and writing codes

  • The box area with a hint of blue as the background color is where you can write code (hereafter referred to as the “code area”).
  • Hit the “Run Code” button to execute all the code inside the code area.
  • You can evaluate (run) code selectively by highlighting the parts you want to run and hitting Command + Enter for Mac (Ctrl + Enter for Windows).
  • If you want to run the codes on your computer, you can first click on the icon with two sheets of paper stacked on top of each other (top right corner of the code chunk), which copies the code in the code area. You can then paste it onto your computer.
  • You can click on the reload button (top right corner of the code chunk, left to the copy button) to revert back to the original code.

Basics

  • This lecture does NOT provide a complete treatment of the basics of the ggplot2 package.

  • Rather, it provides the minimal knowledge of the package so that readers who are not familiar with the package can still keep up with the lecture on map creation.

  • ggplot2 is a general and extensive data visualization tool. It is very popular among R users due to its elegance in and ease of use in generating high-quality figures.

  • It is designed following the “grammar of graphics,”” which makes it possible to visualize data in an easy and consistent manner irrespective of the type of figures generated, whether it is a simple scatter plot or a complicated map.

  • This means that learning the basics of how ggplot2 works directly helps in creating maps as well. This chapter goes over the basics of how ggplot2 works in general.

We use the mpg data to create a simple scatter plot. Here is what mpg dataset looks like:

In ggplot2, you first specify what data to use. The following code declares to R that we will be using mpg as the data for this figure.

Yes, it is a blank canvas. This makes sense because you have not told R how to use the data for visualization.

  • Now that you have specified the data for R to use, we are ready to explain how to use it for visualization.

  • You can achieve this using one of the geom_*() functions available in the ggplot2 package. Here is a short list of some commonly used ones:

    • geom_point(): scatter plot
    • geom_line()” line plot
    • geom_histogram(): histogram
    • geom_boxplot(): box plot
    • geom_sf(): map

Here, let’s create a scatter plot.

Note here that you added a layer defined by geom_point(aes(x = displ, y = hwy)) to g_base.

Let’s now look inside of what is happening in geom_point().

In aes(), x = displ and y = hwy tells R that we want displ on the x-axis and hwy on the y-axis.

Note that different geom_*()s accept/require different options. For example, geom_histogram does not have y as the y-axis is always count.

What happens if we remove aes(). It does not seem to be doing anything. Why can’t we just do this?

Yes, aes() was used to tell R to look for variables inside the data you have specified for R to use earlier in ggplot(data = mpg). Without aes(), R looks for an object named displ (and hwy), which is only defined inside of mpg, thus resulting in the error.

  • Inside a geom_*(), you can specify a number of options to make the figure look different.
  • Different geom_*()s accept different options.
  • Same option names mean different things depending on geom_*() type
  • color: color of the points
  • shape: shape of the points
  • size: size of the points
  • color: color of the borders of the bars
  • fill: color of the inside of the bars
  • shape: no effect
  • linewidth: width of the borders of the bars
  • color: color of the line
  • fill: no effect
  • shape: no effect
  • linewidth: width of the line

Multiple layers

You can easily have multiple layers in a single figure just simply adding geom_() on top of the previous one.

Let’s now add a line plot layer to this.

The way we declared the dataset to use with ggplot(data = mpg) tells R that mpg will be used for every single subsequent geom_*()s unless otherwise specified.

In the code below, mpg is used for both geom_point() and geom_line().

Alternatively, you could specify the dataset locally inside a geom_*() like below, resulting in the same figure as above.


Now, remove data = mpg from the geom_line() above and see what happens. It will result in an error because dataset is not declared either in ggplot() or geom_line(). geom_line() does not know what dataset to use.

With this behavior understood, it is not hard to use multiple datasets in a single figure.

You might have noticed that the line plot looks a bit weird. That is because there are multiple distinct values of hwy observed at the same value of dspl. Let’s get the average value of hwy conditional on displ.


Let’s plot now,

Variable-dependent aesthetics and faceting

  • In the previous examples, all the points and lines had the same color. But, you can use different colors based on the value of a variable.

  • To do so, you need to have the option inside aes().

This code change the color of the points based on the value of model variable.

This code change the shape of the points based on the value of model variable.

This code change the type and color of the lines based on the value of cyl variable.

  • Instead of placing all the information within a single plot, it might be better to have separate panels.
  • Faceted figures are made by effectively splitting the data into groups by a categorical or discrete variable and then apply exactly the same aesthetics to each of the groups.
  • You can use either facet_wrap() or facet_grid() to achieve this.
  • Variable by which you facet needs to be discrete.

Syntax:

ggplot_object +
  facet_wrap(var_1 ~ var_2)
  • var_1: categorical (discrete) variable by which figures are faceted
  • var_2: categorical (discrete) variable by which figures are faceted

You can facet by up to two variables. If you want to facet by only one variable, then put . in place.

Syntax:

ggplot_object +
  facet_grid(var_1 ~ var_2)
  • var_1: categorical (discrete) variable by which figures are faceted (row)
  • var_2: categorical (discrete) variable by which figures are faceted (column)

You can facet by up to two variables. If you want to facet by only one variable, then put . in place.

Yes, it is basically the same, but the order of var_1 and var_2 matters more than facet_wrap() as you will see later.

  • Notice that “var_1” part of the syntax is ., and “var_2” part is factor(cyl)
  • By default, the value of faceting variable (here, cyl) are printed within a strip.

Note

Switch . and factor(cyl) and see what happens.

  • Faceted by trans and factor(cyl).
  • You can decide how many columns or rows the figure should have with ncol and nrow, respectively.

Note

Switch . and factor(cyl) and see what happens.

  • Faceted by trans and factor(cyl).
  • You cannot decide how many columns or rows the figure should have unlike facet_wrap() as the number of levels for the faceting variables dictates them.