`ggplot2`

A layered grammer of graphics in R

Iain Moodie

Expectations

I will try to:
- give you a grammer in which you can describe statistical graphics
- show how this was used when designing ggplot2
- walkthrough some examples (not exhaustive)
- demonstrate some common pitfalls

Expectations

I will not:
- tell you what is a good or bad graphic
- cover everything in ggplot2

Expectations

I expect you to:
- have attended the previous talk about tidyverse
- know now what these symbols mean: |> or %>%
- unless it’s directly related to a slide, save questions to the end

Introduction

What is a (statistical) graphic?

How can we succinctly describe a graphic?
And how can we create the graphic that we have described?
One solution: develop a grammer of graphics

A grammer of graphics

A grammer is “the fundamental principles or rules of an art or science” (OED Online 1989)
Provides the foundation for understanding, describing, and creating graphics
Language analogy:
- Good grammar is just the first step in creating a good sentence

The grammer of graphics

Developed by Wilkinson (2013)
Tweaked and implemented in R by Wickham (2016)

How to build a plot

Scales

Geometries

Aesthetics

Data

A	B	C
1	4	p
2	3	p
3	2	q
4	1	q

Data

simple_data <- tribble(
  ~A, ~B, ~C,
  1, 4, "p",
  2, 3, "p",
  3, 2, "q",
  4, 1, "q"
)

simple_data

# A tibble: 4 × 3
      A     B C    
  <dbl> <dbl> <chr>
1     1     4 p    
2     2     3 p    
3     3     2 q    
4     4     1 q

Mapping data to aesthetics

A	B	C
1	4	p
2	3	p
3	2	q
4	1	q

x	y
1	4
2	3
3	2
4	1

Mapping data to aesthetics

ggplot(
  data = simple_data,
  mapping = aes(x = A, y = B)
  )

Geometries

ggplot(
  data = simple_data,
  mapping = aes(x = A, y = B)
  ) +
  geom_point()

Geometries

ggplot(
  data = simple_data,
  mapping = aes(x = A, y = B)
  ) +
  geom_line()

Scales

A	B	C
1	4	p
2	3	p
3	2	q
4	1	q

x	y	shape
1	4	p
2	3	p
3	2	q
4	1	q

Scales

ggplot(
  data = simple_data,
  mapping = aes(x = A, y = B, shape = C)
  ) +
  geom_point(size = 4)

Scales

ggplot(
  data = simple_data,
  mapping = aes(x = A, y = B, shape = C)
  ) +
  geom_point(size = 4) +
  scale_shape_manual(values = c("p" = 16, "q" = 15))

How to build a plot

Scales

Geometries

Aesthetics

Data

Data: palmer penguins

glimpse(penguins)

Rows: 344
Columns: 8
$ species           <fct> Adelie, Adelie, Adelie, Adelie, Adelie, Adelie, Adel…
$ island            <fct> Torgersen, Torgersen, Torgersen, Torgersen, Torgerse…
$ bill_length_mm    <dbl> 39.1, 39.5, 40.3, NA, 36.7, 39.3, 38.9, 39.2, 34.1, …
$ bill_depth_mm     <dbl> 18.7, 17.4, 18.0, NA, 19.3, 20.6, 17.8, 19.6, 18.1, …
$ flipper_length_mm <int> 181, 186, 195, NA, 193, 190, 181, 195, 193, 190, 186…
$ body_mass_g       <int> 3750, 3800, 3250, NA, 3450, 3650, 3625, 4675, 3475, …
$ sex               <fct> male, female, female, NA, female, male, female, male…
$ year              <int> 2007, 2007, 2007, 2007, 2007, 2007, 2007, 2007, 2007…

How to describe a plot

Aesthetic mapping:
- x = body mass
- y = bill length
- colour = species
Geometries
- points
Scales
- discrete colour = species

`ggplot2` in action

`ggplot2`

Layered grammer of graphics
aesthetic values
- aes()
geometric objects
- geom_*()

`ggplot2`

mapping & aes()

x
y
fill
colour
shape
size
etc

geom_*

geom_point()
geom_line()
geom_bar()
geom_histogram()
geom_boxplot()
etc

Installing and loading ggplot2

ggplot2 is installed with the tidyverse meta package

R console

install.packages("tidyverse")
library(tidyverse)

Or you can install on its own:

R console

install.packages("ggplot2")
library(ggplot2)

ggplot()

ggplot(data = penguins)

ggplot(
  data = penguins,
  mapping = aes(x = flipper_length_mm, y = body_mass_g)
  )

ggplot(
  data = penguins,
  mapping = aes(x = flipper_length_mm, y = body_mass_g)
  ) +
  geom_point()

ggplot(
  data = penguins,
  mapping = aes(x = flipper_length_mm, y = body_mass_g, colour = species)
  ) +
  geom_point()

Two ways to code:

Long form

ggplot(
  data = penguins,
  mapping = aes(x = flipper_length_mm, y = body_mass_g, colour = species)
  ) +
  geom_point()

Condensed form with pipe

penguins |>
  ggplot(aes(x = flipper_length_mm, y = body_mass_g, colour = species)) +
  geom_point()

penguins |>
  filter(species != "Gentoo") |>
  ggplot(aes(x = flipper_length_mm, y = body_mass_g, colour = species)) +
  geom_point()

penguins |>
  filter(species == "Adelie") |>
  ggplot(aes(x = flipper_length_mm, y = body_mass_g, colour = species)) +
  geom_point()

penguins |>
  filter(flipper_length_mm > 190) |>
  ggplot(aes(x = flipper_length_mm, y = body_mass_g, colour = species)) +
  geom_point()

penguins |>
  ggplot(aes(x = flipper_length_mm, y = body_mass_g, colour = species)) +
  geom_point()

penguins |>
  ggplot(aes(x = flipper_length_mm, y = body_mass_g, colour = species)) +
  geom_point() +
  scale_colour_manual(values = c(Gentoo = '#ff8301', Adelie = '#bf5ccb', Chinstrap ='#057076'))

penguins |>
  ggplot(aes(x = flipper_length_mm, y = body_mass_g, colour = species)) +
  geom_point() +
  scale_colour_manual(values = c(Gentoo = '#ff8301', Adelie = '#bf5ccb', Chinstrap ='#057076')) +
  labs(x  = "Flipper length (mm)", y = "Body mass (g)")

penguins |>
  ggplot(aes(x = flipper_length_mm, y = body_mass_g, colour = species)) +
  geom_point() +
  scale_colour_manual(values = c(Gentoo = '#ff8301', Adelie = '#bf5ccb', Chinstrap ='#057076')) +
  labs(x  = "Flipper length (mm)", y = "Body mass (g)") +
  theme_classic()

When and where to use `aes()`

penguins |>
  ggplot(aes(x = flipper_length_mm, y = body_mass_g, colour = species)) +
  geom_point()

When and where to use `aes()`

penguins |>
  ggplot() +
  geom_point(aes(x = flipper_length_mm, y = body_mass_g, colour = species))

When and where to use `aes()`

penguins |>
  ggplot(aes(x = flipper_length_mm, y = body_mass_g, colour = species)) +
  geom_point() +
  geom_smooth(method = "lm")

When and where to use `aes()`

penguins |>
  ggplot(aes(x = flipper_length_mm, y = body_mass_g)) +
  geom_point(aes(colour = species)) +
  geom_smooth(method = "lm")

When and where to use `aes()`

penguins |>
  ggplot(aes(x = flipper_length_mm, y = body_mass_g, colour = species)) +
  geom_point()

When and where to use `aes()`

penguins |>
  ggplot(aes(x = flipper_length_mm, y = body_mass_g, colour = species)) +
  geom_point(aes(size = 1))

When and where to use `aes()`

penguins |>
  ggplot(aes(x = flipper_length_mm, y = body_mass_g, colour = species)) +
  geom_point(aes(size = 2))

When and where to use `aes()`

penguins |>
  ggplot(aes(x = flipper_length_mm, y = body_mass_g, colour = species)) +
  geom_point(aes(size = 3))

When and where to use `aes()`

penguins |>
  ggplot(aes(x = flipper_length_mm, y = body_mass_g, colour = species)) +
  geom_point(size = 3)

When and where to use `aes()`

penguins |>
  ggplot(aes(x = flipper_length_mm, y = body_mass_g, colour = species)) +
  geom_point(size = 1)

When and where to use `aes()`

penguins |>
  ggplot(aes(x = flipper_length_mm, y = body_mass_g, colour = species)) +
  geom_point(size = 5)

When and where to use `aes()`

penguins |>
  ggplot(aes(x = flipper_length_mm, y = body_mass_g, colour = species)) +
  geom_point(size = 5, alpha = 0.5)

When and where to use `aes()`

Mapping data: inside aes()
Not mapping data (just style): outside aes()

Export plots with `ggsave()`

For example

ggsave(filename = "my_plot.pdf")

ggsave(filename = "my_plot.png", width = 6, height = 6, units = "cm", dpi = 300)

by default, will save the last plot made
can change width, height, dpi, etc

Theme

Coordinates

Transformations

Scales

Geometries

Aesthetics

Data

`ggplot2`

aes()
geom_*
scale_*
facet_*
theme_* and theme()

`ggplot2`

use |> or %>% with dplyr verbs first
- filter()
- mutate()
- summarise()
add layers with +
- layers are plotted in order
save your plot with ggsave()

Discussion / questions

References

Wickham, H. 2016. Ggplot2: Elegant graphics for data analysis. Springer-Verlag New York.

Wilkinson, L. 2013. The Grammar of Graphics. Springer Science & Business Media.

ggplot2

Expectations

Expectations

Expectations

Introduction

What is a (statistical) graphic?

A grammer of graphics

The grammer of graphics

How to build a plot

How to build a plot

Data

Data

Mapping data to aesthetics

Mapping data to aesthetics

Geometries

Geometries

Scales

Scales

Scales

How to build a plot

Data: palmer penguins

Data: palmer penguins

How to describe a plot

ggplot2 in action

ggplot2

ggplot2

When and where to use aes()

When and where to use aes()

When and where to use aes()

When and where to use aes()

When and where to use aes()

When and where to use aes()

When and where to use aes()

When and where to use aes()

When and where to use aes()

When and where to use aes()

When and where to use aes()

When and where to use aes()

When and where to use aes()

Facets

Export plots with ggsave()

ggplot2

ggplot2

Discussion / questions

References

`ggplot2`

`ggplot2` in action

`ggplot2`

`ggplot2`

When and where to use `aes()`

When and where to use `aes()`

When and where to use `aes()`

When and where to use `aes()`

When and where to use `aes()`

When and where to use `aes()`

When and where to use `aes()`

When and where to use `aes()`

When and where to use `aes()`

When and where to use `aes()`

When and where to use `aes()`

When and where to use `aes()`

When and where to use `aes()`

Export plots with `ggsave()`

`ggplot2`

`ggplot2`