Data visualization – exercise solutions

Day 4, C

Create your own theme

Load the ggplot2 package and customize your own theme. Look at the built-in themes and the ggthemes package for inspiration.

Tips

1. You can save your customization using the theme(), but that will only modify the current theme

library(ggplot2)
library(palmerpenguins)

mytheme <- theme(strip.background = element_rect(fill = "steelblue"), 
        text = element_text(family = "Comic sans"), 
        plot.background = element_rect(fill = "grey81"), 
        legend.background = element_rect(fill = NA), 
        legend.position = "bottom"
        ) 

ggplot(penguins, aes(x = body_mass_g, y = bill_length_mm, color = species)) + 
  geom_point() + facet_wrap(~ year) + mytheme

Warning: Removed 2 rows containing missing values (geom_point).

theme_set(theme_bw())
ggplot(penguins, aes(x = body_mass_g, y = bill_length_mm, color = species)) + 
  geom_point() + facet_wrap(~ year) + mytheme

Warning: Removed 2 rows containing missing values (geom_point).

2. To make a fully custom theme, start with an existing one, and modify it

my_fulltheme <- theme_grey() + mytheme

ggplot(penguins, aes(x = body_mass_g, y = bill_length_mm, color = species)) + 
  geom_point() + facet_wrap(~ year) + my_fulltheme

Warning: Removed 2 rows containing missing values (geom_point).

3. Save your favorite color scales as a function for easy reuse. Use discrete_scale or continuous_scale.

my_qual_scale <- function(...) {
  discrete_scale("color", scale_name = "OI", 
                 palette = function(x) {
                   res <- palette.colors(x, "Okabe-Ito")[1:x]
                   names(res) <- NULL
                   res
                 }, ...)
}

ggplot(penguins, aes(x = body_mass_g, y = bill_length_mm, color = species)) + 
  geom_point() + facet_wrap(~ year) + my_fulltheme + my_qual_scale()

Warning: Removed 2 rows containing missing values (geom_point).

Complex figures with base graphics

Check out this violin plot

ggplot(penguins, aes(x = species, y = body_mass_g)) + geom_violin() + geom_jitter()

Warning: Removed 2 rows containing non-finite values (stat_ydensity).

Warning: Removed 2 rows containing missing values (geom_point).

How would you produce a similar plot using base graphics? After doing so, reflect, do you appreciate to convenience and flexibility of ggplot2?

Use the density function to calculate the densities by species. The results are stored in a list with elements x and y. Then transform by rotating 90 degrees and mirroring. Plot the result using the polygon function.

dens_by_species <- split(penguins$body_mass_g, penguins$species) |> 
  lapply(density, na.rm = TRUE, cut = 0)

plot(body_mass_g ~ jitter(as.numeric(as.factor(species))), 
     data = penguins, xaxt = "n", xlab = "species")

for(i in 1:length(dens_by_species)) {
  
  den <- dens_by_species[[i]]
  newx <- c(0.2 * (den$y / max(den$y)) + i, 
            rev(-0.2 * (den$y / max(den$y)) + i))
  newy <- c(den$x, rev(den$x))
  
  polygon(newx,  newy)
}
axis(side = 1, at = 1:3, labels = levels(as.factor(penguins$species)))