Presentation 3: ggplot2

Want to code along? If you haven’t already, go to the Data tab of the website and press the DOWNLOAD PRESENTATIONS button. This is presentation3.

Importing libraries and data

library(readxl)
library(writexl)
library(tidyverse)

Load data

In the examples below, we use the downloads dataset. The dataset is available as an Excel file, which we import using the readxl package. Because ggplot2 is part of the tidyverse, we also load the tidyverse packages so we can use its tools for data wrangling and visualization.

We read the downloads dataset from the Excel file, and assign it to a tibble named downloads.

downloads <- read_excel("../Data/downloads.xlsx") %>% 
  filter(size > 0)
downloads

# A tibble: 36,708 × 6
   machineName userID  size  time date                month  
   <chr>        <dbl> <dbl> <dbl> <dttm>              <chr>  
 1 cs18        146579  2464 0.493 1995-04-24 00:00:00 1995-04
 2 cs18        995988  7745 0.326 1995-04-24 00:00:00 1995-04
 3 cs18        317649  6727 0.314 1995-04-24 00:00:00 1995-04
 4 cs18        748501 13049 0.583 1995-04-24 00:00:00 1995-04
 5 cs18        955815   356 0.259 1995-04-24 00:00:00 1995-04
 6 cs18        596819 15063 0.336 1995-04-24 00:00:00 1995-04
 7 cs18        169424  2548 0.285 1995-04-24 00:00:00 1995-04
 8 cs18        386686  1932 0.286 1995-04-24 00:00:00 1995-04
 9 cs18        783767  7294 0.397 1995-04-24 00:00:00 1995-04
10 cs18        788633  4470 3.41  1995-04-24 00:00:00 1995-04
# ℹ 36,698 more rows

Here we have filtered out rows where size is zero, as they do not represent actual downloads and would distort the visualization.

ggplot2: The basic concepts

A ggplot object is a structured description of a plot. The function ggplot() creates the equivalent of a blank sheet of paper, ready for layers to be added.

The two most important components of any ggplot are:

The dataset — typically a data.frame or tibble containing the variables you want to plot
The aesthetics (aes()) — a set of mappings that tell ggplot how to use the variables: x-axis, y-axis and whether to use color, fill, group, size, shape, etc. to represent groups or magnitudes

ggplot(
  downloads,          # dataframe 
  aes(x=machineName)  # x-value
  )

This creates an empty coordinate system with axes defined — essentially a blank canvas.

At this point, ggplot has only drawn the axes and labels because we have not yet specified how the data should be visualized.

The next step is to add a geom layer (for example, geom_bar()) to actually display the data.

Plotting One Variable

A bar chart is created with geom_bar() and is useful for visualizing the distribution of a single categorical variable — for example, machineName in this dataset.

geom_bar() automatically counts the number of observations in each category and uses those counts as the bar heights. By default, categories on the x-axis are displayed in alphabetical order.

ggplot(
  downloads,              # dataframe 
  aes(x=machineName)) +   # categorical variable on x-axis
  geom_bar()              # geometrical object - bar chart

Or we can plot a histogram. It shows the distribution of a continuous variable by dividing the data into bins and counting observations in each bin.

ggplot(
  downloads,               # dataframe 
  aes(x = log2(size))) +   # continuous variable (log2-transformed)
  geom_histogram()         # geometrical object - histogram

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

Aesthetics in ggplot2

In ggplot2, aesthetics (aes()) map variables in your data to visual properties of the plot. In addition to x and y, you can map optional aesthetics such as fill, colour, size, alpha, group, or shape. By changing which aesthetic a variable is mapped to, we can tell ggplot how to display differences between categories.

ggplot(
  downloads,
  aes(x = machineName,
      fill = month)) +     # fill by catergorical variable
  geom_bar()

Here, mapping fill = month tells ggplot to calculate counts separately for each month and assigns a different fill color to each month and stack the bars.

ggplot(
  downloads,
  aes(x = machineName,
      colour = month)) +   # Colour by catergorical variable
  geom_bar()

Here, colour = month only changes the outline color of the bars — the interior remains unfilled (or transparent).

Geometric objects

Geoms (geometric objects) are the core building blocks of ggplot2. They tell ggplot what kind of plot layer to draw. Each geom requires data and aesthetics specified inside aes().
Geoms inherit global aesthetics from the main ggplot() call, but you can override them locally inside each geom if needed.

You can adjust how geoms are displayed using various options. For bar charts, position adjustments control how overlapping or grouped data appear. In the previous example, mapping fill, colour, or group = month subdivided each bar by month, producing stacked bars (the default).

You can change this behavior by setting the position argument inside geom_bar():

position = "dodge" → places bars for each group side by side
position = "fill" → normalizes bar heights so they represent proportions (percentages).

ggplot(
  downloads,
  aes(x = machineName,
      fill = month)) + 
  geom_bar(position = "dodge")  # side by side

ggplot(
  downloads,
  aes(x = machineName,
      fill = month)) + 
  geom_bar(position = "fill")   # proportions

Plotting Two Variable - A simple Column Chart

To make a bar chart from pre-summarized data, we use geom_col() and it requires both x and y values.

Suppose we want to plot the total size of all downloads per machine.
We first create a summary tibble dl_sizes using:

group_by() to group the dataset by machineName
summarise() to calculate the sum of size within each group

In the example below, we also rescale the download sizes from bytes to megabytes by dividing by 1,000,000.

dl_sizes <- downloads %>% 
  group_by(machineName) %>% 
  summarise(size = sum(size))

We now plot machineName on the x-axis and the total size (in MB) on the y-axis:

ggplot(
  dl_sizes, 
  aes(x = machineName, 
      y = size/10^6)) + 
  geom_col()

A Column Chart with Ordered Bars

By default, categorical variables in R are displayed in alphabetical order on the x-axis. Suppose we want the machines in our bar chart to be ordered by increasing total download size instead. This is actually not that easy in R.
We can do this by converting machineName into a factor with levels ordered according to download size.

First, arrange the dl_sizes tibble by size using tidyverse function arrange.

dl_sizes <- dl_sizes %>% 
  arrange(size)
dl_sizes

# A tibble: 5 × 2
  machineName      size
  <chr>           <dbl>
1 pluto        72605544
2 cs18        100593281
3 tweetie     104379794
4 piglet      158149841
5 kermit      175032552

Next, set the levels of machineName according to the arranged order:

dl_sizes <- dl_sizes %>% 
  mutate(machineName = factor(machineName, levels = dl_sizes$machineName))

dl_sizes$machineName

[1] pluto   cs18    tweetie piglet  kermit 
Levels: pluto cs18 tweetie piglet kermit

Now machineName is a factor with levels ordered from lowest to highest total download size.

Finally, use the same ggplot() code as before to make the plot — now the bars will be ordered by total size rather than alphabetically:

ggplot(
  dl_sizes, 
  aes(x = machineName, 
      y = size/10^6)) + 
  geom_col()

A box plot

Boxplots are great to get an overview of continues variables, identifying outliers and comparing distributions across categories.

Here we will for the first time try to save the syntaxical description of a plot in an R variable. This makes it easy to reuse and modify the same plot - for example, to change axes scales or colors - without rewriting the entire code. Let’s try this out and create a boxplot of monthly download sizes for each machine.

Here, we store the bar chart in a variable called p. Running this code does not display the plot yet, but a variable named p now appears in the Global Environment.

p <-                       # save in object p
  ggplot(                  # call ggplot
    downloads,             # data
    aes(x = machineName,   # x and y and fill
        y = size, 
        fill = machineName)) +
  geom_boxplot()           # boxplot

Because the data are skewed, we use a logarithmic y-scale to make the plot more interpretable. A logarithmic scale is useful when the data are highly skewed or span several orders of magnitude.

p <- p + scale_y_log10()          # Change the scale of the y-axis:

Additionally, themes and titles can be added as a layer to any ggplot if you prefer a theme other than the default grey background.

Because the plot is saved in p, we can easily modify it by adding more layers:

plot1 <- p + 
  theme_minimal() +         # add theme ..................
    # theme_bw()
    # theme_classic()
    # theme_dark()
  labs(x = 'Machines',     # add label ..................
       y = 'Size in Mb',
       title = 'Boxplot')
plot1

Violin plot with `geom_violin`

A violin plot shows the distribution of a continuous variable across different categories, combining the features of a box plot and a density plot.

p <-                         # save in object p
  ggplot(                    # call ggplot
    downloads,               # data
    aes(x = machineName,     # x and y and fill
        y = size, 
        fill = machineName)) +
  geom_violin() +            # draw violin plots
  scale_y_log10()            # scale
p

We can enhance the plot by overlaying individual data points (with jittering to avoid overplotting), using a minimal theme, and customizing the fill colors:

plot2 <- p + 
  geom_jitter(           # overlay geom
    size = 0.1, alpha = 0.05) + 
  theme_minimal() +      # theme
  scale_fill_manual(
    values=c("#2A2D43",  # change colors
             "#91A6FF", 
             "#C7EDE4", 
             "#AFA060", 
             "#AD8350"))
plot2

Scatter plot with `geom_point` and Faceting

A scatter plot is used to explore the relationship between two numerical variables by plotting them on the x- and y-axes. You can visualize a third variable (or a condition) by mapping it to the size ,shape or color aesthetic.

First, we plot download size vs. time using geom_point() to draw individual data points. We map time > 500 to the size aesthetic so downloads taking longer than 500 appear larger. We also map log10(time) to color (so color shows smooth continuous gradient), and use shape to distinguish machineName. Because both variables are skewed, we use log scales on both axes with scale_x_log10() and scale_y_log10().

plot3 <-                       # save in object p
  ggplot(                  # call ggplot
    downloads,             # data
    aes(x = size,          # x and y and color
        y = time, 
        size = time > 500,
        color = log10(time),
        shape = machineName)) +   # add size, color and shape by machineName 
  geom_point() +          # Scaterplot
  scale_colour_continuous(palette = c("darkblue", "gold")) +
  scale_x_log10() +       # scale x and y
  scale_y_log10() + 
  theme_minimal()         # theme

plot3

Warning: Using size for a discrete variable is not advised.

Sometimes adding several aesthetics (like both color and shape) makes a plot too busy, especially if there are many categories. Instead, we can use faceting to split the data into multiple smaller plots - one for each category - making it easier to compare groups.

#plot3 + facet_grid(vars(month))
plot3 + facet_wrap(vars(machineName))

Warning: Using size for a discrete variable is not advised.

This wraps the plots into multiple rows and columns automatically, which is helpful when there are many facet levels.

Now you have space to highlight important patterns without overcrowding the plot. For example, faceting makes it easier to see the relationship (correlation) between download size and time within each machine, and to spot extreme cases. In particular, we can highlight very long downloads (e.g. time > 500) by making those points larger, so they stand out clearly.

Daily summary statistics

Next we want to visualize the number of downloads done each day for each of the 6 machines. We compute the cumulated number of downloads within each of the 6 machines over the days. As before we do this via the following steps:

Using group_by() we group the dataset by both machineName and date.
Using summarize() we count the number for within each machine and date.
Using mutate() we cumulate the number of downloads over the dates within the machines.

daily_downloads <- 
  downloads %>%
    group_by(machineName, date) %>% 
    summarize(dl_count = n()) %>%
    mutate(total_dl_count = cumsum(dl_count))

`summarise()` has grouped output by 'machineName'. You can override using the
`.groups` argument.

daily_downloads

# A tibble: 337 × 4
# Groups:   machineName [5]
   machineName date                dl_count total_dl_count
   <chr>       <dttm>                 <int>          <int>
 1 cs18        1994-11-22 00:00:00      168            168
 2 cs18        1994-11-23 00:00:00      191            359
 3 cs18        1994-11-24 00:00:00      256            615
 4 cs18        1994-11-25 00:00:00       13            628
 5 cs18        1994-11-28 00:00:00      133            761
 6 cs18        1994-11-29 00:00:00        8            769
 7 cs18        1994-11-30 00:00:00       14            783
 8 cs18        1994-12-01 00:00:00       35            818
 9 cs18        1994-12-02 00:00:00       61            879
10 cs18        1994-12-03 00:00:00       11            890
# ℹ 327 more rows

Line plot with `geom_line`

The geom called geom_line() is used to insert lines, which is ideal for showing trends over time or continuous changes. In the code below you see a first attempt to write R code that visualizes the cumulated total download size over the dates.

We create a plot of the cumulative total download count over time:

ggplot(                      # call ggplot
  daily_downloads,           # data
  aes(x = date,              # x and y
      y = total_dl_count)) + # line
  geom_line()

If we want a separate line for each machine, we must tell ggplot how to group the data. We can do this by mapping color = machineName (and optionally color by machine as well):

plot4 <- ggplot(
  daily_downloads, 
  aes(x = date, 
      y = total_dl_count, 
      colour = machineName)) + 
  geom_line() +
  theme_minimal()     # theme
plot4

Arrange multiple ggplots on the same page

In R there is a whole infrastructure of extension packages built around ggplot, that facilitate different types of plot modifications, arrangements, and animation.

R package ggpubr is one of them and it helps to produce publication-ready plots using ggplot2.

# install.packages("ggpubr")
library(ggpubr)

figure <- ggarrange(plot1, plot2, plot3, plot4,
                   align = "hv",
                    labels = c ("A", "B", "C", "D"),
                   ncol = 2, nrow = 2, 
                   legend = "none")

Warning: Using size for a discrete variable is not advised.

figure

## Save Tibble and Save Plot:

#write_xlsx(daily_downloads, "daily_downloads.xlsx")

#ggsave(filename = "vionlinPlot.pdf", plot = p2, height = 5, width = 10)

#pdf("vionlinPlot2.pdf", height = 5, width = 10)
#p2
#dev.off()

Conclusion

The ggplot2 package is a powerful and versatile tool for making plots.
We think, that the generated plots are beautiful.
As far as we know some plots, e.g. using faceting (see the exercises), in practice are only producible in ggplot2.
Learning the syntax needed for making specific plots is a challenge. The best way (the only way!?) to learn is to practice. You may start by solving the exercise sheet. After you have gotten used to the basic ideas you can find a lot of help on the internet. We remark, that there have been several updates of the ggplot2 package over the recent years. And some of the old entries that might pop up when you google a ggplot2-issue might be outdated.
Not all things can be made in ggplot2! For a geometrical object to be available it need to have a syntaxical description, and it needs to be implemented. Other things require computer-hacks to be made (e.g. using different orderings of categorical variables in faceted plots).