Import the data

The mpg dataset is included as part of the ggplot2 library:

library(ggplot2)
data("mpg")

mpg

## # A tibble: 234 x 11
##    manufacturer model    displ  year   cyl trans   drv     cty   hwy fl   
##    <chr>        <chr>    <dbl> <int> <int> <chr>   <chr> <int> <int> <chr>
##  1 audi         a4        1.80  1999     4 auto(l… f        18    29 p    
##  2 audi         a4        1.80  1999     4 manual… f        21    29 p    
##  3 audi         a4        2.00  2008     4 manual… f        20    31 p    
##  4 audi         a4        2.00  2008     4 auto(a… f        21    30 p    
##  5 audi         a4        2.80  1999     6 auto(l… f        16    26 p    
##  6 audi         a4        2.80  1999     6 manual… f        18    26 p    
##  7 audi         a4        3.10  2008     6 auto(a… f        18    27 p    
##  8 audi         a4 quat…  1.80  1999     4 manual… 4        18    26 p    
##  9 audi         a4 quat…  1.80  1999     4 auto(l… 4        16    25 p    
## 10 audi         a4 quat…  2.00  2008     4 manual… 4        20    28 p    
## # ... with 224 more rows, and 1 more variable: class <chr>

glimpse(mpg)

## Observations: 234
## Variables: 11
## $ manufacturer <chr> "audi", "audi", "audi", "audi", "audi", "audi", "...
## $ model        <chr> "a4", "a4", "a4", "a4", "a4", "a4", "a4", "a4 qua...
## $ displ        <dbl> 1.8, 1.8, 2.0, 2.0, 2.8, 2.8, 3.1, 1.8, 1.8, 2.0,...
## $ year         <int> 1999, 1999, 2008, 2008, 1999, 1999, 2008, 1999, 1...
## $ cyl          <int> 4, 4, 4, 4, 6, 6, 6, 4, 4, 4, 4, 6, 6, 6, 6, 6, 6...
## $ trans        <chr> "auto(l5)", "manual(m5)", "manual(m6)", "auto(av)...
## $ drv          <chr> "f", "f", "f", "f", "f", "f", "f", "4", "4", "4",...
## $ cty          <int> 18, 21, 20, 21, 16, 18, 18, 18, 16, 20, 19, 15, 1...
## $ hwy          <int> 29, 29, 31, 30, 26, 26, 27, 26, 25, 28, 27, 25, 2...
## $ fl           <chr> "p", "p", "p", "p", "p", "p", "p", "p", "p", "p",...
## $ class        <chr> "compact", "compact", "compact", "compact", "comp...

Each row represents a model of car sold in a given year.³ hwy identifies the highway miles per gallon for the vehicle.

Histogram

ggplot(mpg, aes(hwy)) +
  geom_histogram()

## `stat_bin()` using `bins = 30`. Pick better value with `binwidth`.

It appears there is a high concentration of vehicles with highway fuel efficiency between 20 and 30 mpg, with a smaller number of vehicles between 15-20 and some outliers with high fuel efficiency (larger values indicate more efficient vehicles). To view the actual data points, we use geom_rug():

ggplot(mpg, aes(hwy)) +
  geom_histogram() +
  geom_rug()

## `stat_bin()` using `bins = 30`. Pick better value with `binwidth`.

One thing geom_rug() does is illustrate that while hwy is a continuous variable, it is measured in integer units - that is, there are no fractional values in the dataset. 26 miles per gallon on the highway is the most common mpg rate in the dataset. Why is that? Something perhaps to investigate further.

By default, geom_histogram() bins the observations into 30 intervals of equal width. You can adjust this using the bins parameter:

ggplot(mpg, aes(hwy)) +
  geom_histogram(bins = 50) +
  geom_rug()

ggplot(mpg, aes(hwy)) +
  geom_histogram(bins = 10) +
  geom_rug()

Notice the spike at 9 micrograms per cubic meter. We cannot immediately tell what causes this spike. but it could be worth exploring later.

Bar chart

ggplot(mpg, aes(class)) +
  geom_bar()

To examine the distribution of a categorical variable, we can use a bar chart. Here we see the most common type of vehicle in the dataset is an SUV, not surprising given Americans’ car culture.

Two-dimensional graphs

Two-dimensional graphs are visualizations that are naturally designed to visualize two variables. For instance, if you have a discrete variable and a continuous variable, you could use a box plot to visualize the distribution of the values of the continuous variable for each category in the discrete variable:

ggplot(mpg, aes(class, hwy)) +
  geom_boxplot()

Here we see that on average, compact and midsize vehicles have the highest highway fuel efficiency whereas pickups and SUVs have the lowest fuel efficiency. What might explain these differences? Another question you could explore after viewing this visualization.

If you have two continuous variables, you may use a scatterplot which maps each variable to an \(x\) or \(y\)-axis coordinate. Here we visualize the relationship between engine displacement (the physical size of the engine) and highway fuel efficiency:

ggplot(mpg, aes(displ, hwy)) +
  geom_point()

As engines get larger, highway fuel efficiency declines.

Multiple window plots

Sometimes you want to compare the conditional distribution of a variable across specific groups or subsets of the data. To do that, we implement a multiple window plot (also known as a trellis or facet graph). This involves drawing the same plot repeatedly, using a separate window for each category defined by a variable. For instance, if we want to examine variation in highway fuel efficiency separately for type of drive (front wheel, rear wheel, or 4 wheel), we could draw a graph like this:

ggplot(mpg, aes(hwy)) +
  geom_histogram() +
  facet_wrap(~ drv)

## `stat_bin()` using `bins = 30`. Pick better value with `binwidth`.

Highway fuel efficiency is right-skewed for 4 and rear wheel drive vehicles, whereas front wheel drive vehicles are generally unskewed with a couple outliers of 40+ mpg.

You may also want to use a multiple windows plot with a two-dimensional graph. For example, the relationship between engine displacement and highway fuel efficiency by drive type:

ggplot(mpg, aes(displ, hwy)) +
  geom_point() +
  facet_wrap(~ drv)

Utilizing additional channels

If you want to visualize three or more dimensions of data without resorting to 3D animations⁴ or window plots, the best approach is to incorporate additional channels into the visualization. Channels are used to encode variables inside of a graphic. For instance, a scatterplot uses vertical and horizontal spatial position channels to encode the values for two variables in a visually intuitive manner.

Depending on the type of graph and variables you wish to encode, there are several different channels you can use to encode additional information. For instance, color can be used to distinguish between classes in a categorical variable.

ggplot(mpg, aes(displ, hwy, color = class)) +
  geom_point()

We can even use a fourth channel to communicate another variable (number of cylinders) by making use of the size channel:

ggplot(mpg, aes(displ, hwy, color = class, size = cyl)) +
  geom_point()

Note that some channels are not always appropriate, even if they can technically be implemented. For example, instead of using a color channel to visualize class, why not distinguish each type of car using the point’s shape?

ggplot(mpg, aes(displ, hwy, shape = class)) +
  geom_point()

## Warning: The shape palette can deal with a maximum of 6 discrete values
## because more than 6 becomes difficult to discriminate; you have 7.
## Consider specifying shapes manually if you must have them.

## Warning: Removed 62 rows containing missing values (geom_point).

With this many categories, shape is not very useful in visually distinguishing between each car’s class.