Visualising and Plotting Data

ASI: Introduction to R

Author

Affiliation

Dr Stevie Pederson

Black Ochre Data Labs
The Kids Research Institute Australia

Visualisation in R

Start a New R Script

• Call the new script: BasicVisualisation.R
• Load our favourite package at the top of the script

library(tidyverse)

• Load the my_penguins dataset

my_penguins <- read_csv("data/my_penguins.csv")

Introducing The Penguins

my_penguins

# A tibble: 333 × 8
   species island    bill_length_mm bill_depth_mm flipper_length_mm body_mass_g
   <chr>   <chr>              <dbl>         <dbl>             <dbl>       <dbl>
 1 Adelie  Torgersen           39.1          18.7               181        3750
 2 Adelie  Torgersen           39.5          17.4               186        3800
 3 Adelie  Torgersen           40.3          18                 195        3250
 4 Adelie  Torgersen           36.7          19.3               193        3450
 5 Adelie  Torgersen           39.3          20.6               190        3650
 6 Adelie  Torgersen           38.9          17.8               181        3625
 7 Adelie  Torgersen           39.2          19.6               195        4675
 8 Adelie  Torgersen           41.1          17.6               182        3200
 9 Adelie  Torgersen           38.6          21.2               191        3800
10 Adelie  Torgersen           34.6          21.1               198        4400
# ℹ 323 more rows
# ℹ 2 more variables: sex <chr>, year <dbl>

• Contains multiple measurements for penguins recorded around Palmer Station, Antarctica
• Slightly modified version from https://allisonhorst.github.io/palmerpenguins/

Base Plotting in R

• R comes with some very powerful plotting capabilities
  • Provided in the base package graphics
  • Always loaded with every session
• Examples are often extremely helpful
• People used happily for decades
  • The release of ggplot2 changed everything
• Let’s quickly explore base plotting before moving to the good stuff

Base Plotting In R

• Simple plots are usually easy
  • Complex figures can get really messy
• Using the cars dataset
  • speed (mph)
  • dist (ft) each car takes to stop

plot(cars)

Base Plotting In R

• The first two columns were automatically placed on the x & y axis
• We could set values for x & y manually
• Automatically decided to plot using points

plot(x = cars$speed, y = cars$dist)

Base Plotting In R

• Using the my_penguins dataset to compare flipper length and body mass

plot(x = my_penguins$body_mass_g, y = my_penguins$flipper_length_mm) 

Base Plotting In R

• The function boxplot() can also create simple figures easily
• For categorical variables (i.e. factors) we use R formula notation
  • y ~ x \(\implies\) y depends on x, or
  • y ~ x \(\implies\) y is a function of x

## Make a simple boxplot showing the weights by species
boxplot(body_mass_g ~ species, data = my_penguins)

• The dependent variable will always appear on the y-axis
• The predictor will always appear on the x-axis

Base Plotting In R

• We can also use combinations of predictor variables

## Separate by species and sex
boxplot(body_mass_g ~ sex + species, data = my_penguins)

Base Plotting In R

• Histograms can be produced on an individual column
  • The number of breaks can be set manually
• The default is pretty useful here
  • Generally simple figures without complexity

hist(my_penguins$body_mass_g, breaks = 20, xlab = "Body Mass (g)")

Visualisation With ggplot2

The Grammar of Graphics

• ggplot2 has become the industry standard for visualisation (Wickham 2016)
• Core & essential part of the tidyverse
• Developed by Hadley Wickham as his PhD thesis
• An implementation of The Grammar of Graphics (Wilkinson 2005)
  • Breaks visualisation into layers
  • Each layer added as a new line of code

The Grammar of Graphics

Taken from https://r.qcbs.ca/workshop03/book-en/grammar-of-graphics-gg-basics.html

Beginning Our Plot

## The initial call to ggplot defines
## the data layer only
ggplot(my_penguins)

Adding Layers

• I’m spreading my code to fit the narrow column in the slide

## Adding aesthetic mappings sets the 
## data range (via the axes)
ggplot(my_penguins) +
  aes(
    x = body_mass_g, 
    y = flipper_length_mm
  )

Adding Layers

• Notice how the call to ggplot() was followed by +
• This tells R: “But wait, there’s more to come!”
• Is how we add the layers described in The Grammar of Graphics

Adding Points

## Adding `geom_point()` now tells
## ggplot what geometry to use
ggplot(my_penguins) +
  aes(
    x = body_mass_g, 
    y = flipper_length_mm
  ) +
  geom_point()

Tidying The Code

• To write more clearly

ggplot(my_penguins) + # 1: Define the data layer
  aes(x = body_mass_g, y = flipper_length_mm) + # 2: Aesthetic mappings
  geom_point() # 3: Geometry

Adding Layers

• I’ll hide my figures from here to the code looks nicer

• Now we can easily add additional layers
  • e.g. a smooth curve as a statistics layer

ggplot(my_penguins) + # 1: Define the data layer
  aes(x = body_mass_g, y = flipper_length_mm) + # 2: Aesthetic mappings
  geom_point() + # 3: Geometry
  stat_smooth() # 5: Statistics

Adding Layers

• Change to a regression line without an error region:
  • method = "lm", se = FALSE

ggplot(my_penguins) + # 1: Define the data layer
  aes(x = body_mass_g, y = flipper_length_mm) + # 2: Aesthetic mappings
  geom_point() + # 3: Geometry
  stat_smooth(method = "lm", se = FALSE) # 5: Statistics

Additional Aesthetics

• We might like to colour points by species
• Adding this to the main aes() will send this mapping to all layers
  • Will produce a regression line for each species

ggplot(my_penguins) + # 1: Define the data layer
  aes(x = body_mass_g, y = flipper_length_mm, colour = species) + # 2: Mappings
  geom_point() +  # 3: Geometry
  stat_smooth(method = "lm", se = FALSE) # 5: Statistics

Additional Aesthetics

• We can also include the sex of penguins on this figure
• Can easily add as an aesthetic mapping

ggplot(my_penguins) + # 1: Define the data layer
  aes(x = body_mass_g, y = flipper_length_mm, colour = species, shape = sex) + 
  geom_point() +  # 3: Geometry
  stat_smooth(method = "lm", se = FALSE) # 5: Statistics

• This doesn’t work well for the regression lines
• Lines don’t have a shape aesthetic \(\implies\) ignored by stat_smooth()

Additional Aesthetics

• Aesthetic mappings set in the first call are passed to all layers
• Aesthetics can also be set as layer-specific

ggplot(my_penguins) + # 1: Define the data layer
  aes(x = body_mass_g, y = flipper_length_mm, colour = species) + # 2: Mappings
  geom_point(aes(shape = sex)) +  # 3: Geometry with layer-specific aes
  stat_smooth(method = "lm", se = FALSE) # 5: Statistics

Facets

• Separate panels are known as facets in ggplot2
  • Plot male & female penguins in separate panels

ggplot(my_penguins) + # 1: Define the data layer
  aes(x = body_mass_g, y = flipper_length_mm, colour = species) + # 2: Mappings
  geom_point(aes(shape = sex)) +  # 3: Geometry with layer-specific aes
  facet_wrap(~sex) + # 4: Facets
  stat_smooth(method = "lm", se = FALSE) # 5: Statistics

Facets

• We can “free” each axis within each panel
  • Not always informative

ggplot(my_penguins) + # 1: Define the data layer
  aes(x = body_mass_g, y = flipper_length_mm, colour = species) + # 2: Mappings
  geom_point(aes(shape = sex)) +  # 3: Geometry with layer-specific aes
  facet_wrap(~sex, scales = "free") + # 4: Facets
  stat_smooth(method = "lm", se = FALSE) # 5: Statistics

Co-ordinates

• An additional layer is the co-ordinate system
  • Defaults are always chosen (unlike geometry)
  • Can modify if we choose
• Quite obvious for points
  • Can transform axes to log₁₀ etc

• Applies to colours and shapes as well
• Set using scale_*() functions

Co-ordinates

• Manually set the point shapes
• The options can be seen using ?pch

ggplot(my_penguins) + # 1: Define the data layer
  aes(x = body_mass_g, y = flipper_length_mm, colour = species) + # 2: Mappings
  geom_point(aes(shape = sex)) +  # 3: Geometry with layer-specific aes
  facet_wrap(~sex, scales = "free") + # 4: Facets
  stat_smooth(method = "lm", se = FALSE) + # 5: Statistics
  scale_shape_manual(values = c(1, 19)) # 6: Co-ordinates (i.e. scales)

Co-ordinates

• Colour scales are also set like this
  • scale_colour_brewer() is an excellent starting point

ggplot(my_penguins) + # 1: Define the data layer
  aes(x = body_mass_g, y = flipper_length_mm, colour = species) + # 2: Mappings
  geom_point(aes(shape = sex)) +  # 3: Geometry with layer-specific aes
  facet_wrap(~sex, scales = "free") + # 4: Facets
  stat_smooth(method = "lm", se = FALSE) + # 5: Statistics
  scale_shape_manual(values = c(1, 19)) + # 6: Co-ordinates (i.e. scales)
  scale_colour_brewer(palette = "Set1")

• Check brewer palettes using RColorBrewer::display.brewer.all()

Co-ordinates

• Multiple options for discrete colour palettes
  • scale_colour_manual(values = c("red", "blue", "green"))
  • scale_colour_viridis_d()

• Colourblind palettes are provided in ggthemes
  • scale_colour_colorblind()

Co-ordinates

library(ggthemes)
ggplot(my_penguins) + # 1: Define the data layer
  aes(x = body_mass_g, y = flipper_length_mm, colour = species) + # 2: Mappings
  geom_point(aes(shape = sex)) +  # 3: Geometry with layer-specific aes
  facet_wrap(~sex, scales = "free") + # 4: Facets
  stat_smooth(method = "lm", se = FALSE) + # 5: Statistics
  scale_shape_manual(values = c(1, 19)) + # 6: Co-ordinates (i.e. scales)
  scale_colour_colorblind() # Stoopid American spelling

• Scales often require an understanding of discrete vs continuous, e.g.
  • scale_x_continuous() for continuous data
  • scale_x_discrete() for categorical data
  • scale_colour_gradient() for continuous data (e.g. heatmaps)

Themes

• The final grammar of graphics layer \(\implies\) themes
• Controls overall appearance:
  • font-sizes, background style, title/legend placement etc
• Will explore in detail later if we have time

• We do need to get rid of the horrible grey background though

Themes

• theme_bw() applies a generally usable set of defaults

library(ggthemes)
ggplot(my_penguins) + # 1: Define the data layer
  aes(x = body_mass_g, y = flipper_length_mm, colour = species) + # 2: Mappings
  geom_point(aes(shape = sex)) +  # 3: Geometry with layer-specific aes
  facet_wrap(~sex, scales = "free") + # 4: Facets
  stat_smooth(method = "lm", se = FALSE) + # 5: Statistics
  scale_shape_manual(values = c(1, 19)) + # 6: Co-ordinates (i.e. scales)
  scale_colour_colorblind() + # Stoopid American spelling
  theme_bw() # 7: The overall theme

• I often call theme_set(theme_bw()) at the start of a session
  • Sets as the theme for all subsequent plots

Additional Geometries

Additional Geometries

• The choice of appropriate geometry is usually data-driven
  • Same principle as aes() / geom_point()
  • May have different mappings

• Lines: geom_line(), geom_abline(), geom_hline(), geom_vline()
• Distributions: geom_boxplot(), geom_violin()
• Histograms: geom_histogram(), geom_density()
• Bar Plots: geom_bar(), geom_col() + geom_errorbar()
• Heatmaps: geom_tile(), geom_raster(), geom_rect()

Making Boxplots

ggplot(
  my_penguins,
  aes(species, body_mass_g)
) +
  geom_boxplot() +
  theme_bw()

Making Boxplots

## Map sex to the fill aesthetic
ggplot(my_penguins) +
  aes(
    species, body_mass_g, fill = sex
  ) +
  geom_boxplot() +
  theme_bw()

Histograms

• The default histogram usually looks terrible \(\implies\) easy to fix

## Map sex to the fill aesthetic
ggplot(my_penguins) +
  aes(x = body_mass_g, fill = sex) +
  geom_histogram() +
  theme_bw()

Histograms

• The default histogram usually looks terrible \(\implies\) easy to fix
  • Changing binwidth and setting colour (outline)

ggplot(my_penguins) +
  aes(x = body_mass_g, fill = sex) +
  geom_histogram(
    colour = "black", binwidth = 100
  ) +
  facet_grid(sex~species) +
  theme_bw()

Closing Comments

• The top-level aes() can also be set in the first call to ggplot()
  • Is personal preference \(\implies\) showed layers clearly today

## My code usually looks more like this
ggplot(
  my_penguins, 
  aes(x = body_mass_g, fill = sex)
) +
  geom_histogram(colour = "black", binwidth = 100) +
  facet_grid(sex~species) +
  theme_bw()

Closing Comments

• Will explore additional geometries through the course
• Also discuss more detailed ggplot2 customisation
• Non ggplot2 options can also be effective:
  • corrplot() from the package corrplot
  • pheatmap() from the package pheatmap
  • Venn Diagrams from VennDiagram
  • UpSet plots from UpSetR (is actually ggplot)

Challenge

1. Load the pigs dataset and create a boxplot
   • Show dose across the x-axis
   • Fill by supplement type
2. Experiment with geom_violin() as an alternative

References

Wickham, Hadley. 2016. Ggplot2: Elegant Graphics for Data Analysis. Springer-Verlag New York. https://ggplot2.tidyverse.org.

Wilkinson, Leland. 2005. The Grammar of Graphics. Springer New York, NY. https://doi.org/https://doi.org/10.1007/0-387-28695-0.