Exercises for session 1.2

library(ggplot2)
library(grid)

# Some graphics devices don't support gradients and patterns. We set the
# device to ragg, which does support these.
knitr::opts_chunk$set(dev = "ragg_png")

Headings

Exercise 1.2.1

We can use the label attribute in columns to automatically label a variable. Complete the chunk below to set a label attribute for the unemploy column in the economics dataset. Per ?economics, the unemploy variable gives the number of unemployed people in thousands.

attr(economics$date, "label") <- "Year"
...

ggplot(economics, aes(date, unemploy)) +
  geom_area(alpha = 0.4, colour = "black")

Exercise 1.2.2

If don’t have pre-set labels, we can also use the labs(dictionary) argument to populate column-label pairs. Use this argument to label the variables in the plot below. You can use ?mpg to find out what the columns are.

ggplot(mpg, aes(displ, hwy, colour = drv)) +
  geom_point() +
  labs(dictionary = ...)

Patterns and gradients

Exercise 1.2.3

Use grid::linearGradient() to set up a horizontal colour gradient. Which arguments do you have to tweak to change the default diagonal gradient to a horizontal one? How would you set a vertical gradient instead?

gradient <- ...

ggplot(economics, aes(date, unemploy)) +
  geom_area(fill = list(gradient), colour = "black")

Exercise 1.2.4

The following code sets up a crosshatch pattern that can be used as the fill aesthetic.

width <- height <- unit(3, "mm")
crosshatch <- pattern(
  segmentsGrob(
    x0 = c(0, 1), x1 = c(1, 0), 
    y0 = c(0, 0), y1 = c(1, 1),
    gp = gg_par(col = "black", lwd = 0.5),
    vp = viewport(width = width, height = height)
  ),
  width = width, height = height,
  extend = "repeat"
)

ggplot(economics, aes(date, unemploy)) +
  geom_area(fill = crosshatch, colour = "black")

Create a new pattern of your liking. Can you design a polka-dot or checkerboard pattern? You can use grid::circleGrob() and grid::rectGrob() for circles and rectangles respectively. We’ll use it to create a manual fill scale of patterns.

my_pattern <- pattern(
  ...
)

patterns <- list(
  crosshatch,
  gradient,
  my_pattern
)

ggplot(palmerpenguins::penguins) +
  aes(species, fill = island) +
  geom_bar() +
  scale_fill_manual(values = patterns)

Delayed evaluation

Exercise 1.2.5

In the plot below, can you redirect the alpha aesthetic to represent the count computed variable? You can set the fill aesthetic to a fixed colour like "black". For an extra challenge, can you use scale_alpha_continuous() to anchor 0 at complete transparency?

ggplot(diamonds) +
  aes(carat, price) +
  stat_bin_2d(
    binwidth = c(0.05, 200)
  )

Polar coordinates

Exercise 1.2.6

Plot p displays a Cartesian bar chart of car manufacturers. Add a coord_radial() with arguments to p to make a half-ring shape coxcomb/windrose chart with nicely displayed text.

p <- ggplot(mpg) +
  aes(manufacturer, label = manufacturer) +
  geom_bar() +
  geom_text(
    aes(y = 37),
    # Prevent duplicated labels
    data = ~ dplyr::filter(.x, !duplicated(manufacturer)),
    angle = 90, hjust = 1
  ) +
  scale_x_discrete(guide = "none")
p

Facets

Exercise 1.2.7

Given the following code, adapt the facet_wrap() statement to:

Mirror the facet order
Place labelled axes at the bottom of every panel
Place axis ticks on all y-axes

ggplot(diamonds) +
  aes(carat, price) +
  geom_point(shape = ".") +
  facet_wrap(~ cut, dir = "lt")