Set a theme and palette for all plots

Question

I'm trying to simplify my plots in ggplot2. Suppose I want to create a scatterplot from iris dataset:

ggplot(iris, aes(x=Petal.Length, y=Petal.Width, colour=Species)) +
  geom_point()

But suppose I don't like ggplot2 default theme and palette. Let's say I want to use theme_bw and Dark2 palette:

ggplot(iris, aes(x=Petal.Length, y=Petal.Width, colour=Species)) +
  geom_point() +
  theme_bw() +
  scale_color_brewer(palette="Dark2")

And suppose I have lots of plots and I want all of them using theme_bw and Dark2 palette. I know I can use theme_set(theme_bw()) to make all my plots have the black and white theme. Is there a similar function to make all my plots use Dark2 palette? In other words, how can I run a code like

ggplot(iris, aes(x=Petal.Length, y=Petal.Width, colour=Species)) +
  geom_point()

and have theme_bw and Dark2 palette in all my plots?

Answer 1

One solution would be to write a custom wrapper:

ggcust <- function(...){
  ggplot(...) +
    theme_bw()
}

fill in all the theme options you need, then use it like this:

ggcust(data = mtcars, aes(x = mpg, y = cyl)) +
  geom_point()

Answer 2

You can also put layers into a list:

gglayer_theme <- list(
  theme_bw(),
  scale_color_brewer(palette="Dark2")
)

And treat the list like a new layer (Note + becomes , in this list notation):

ggplot(iris, aes(x=Petal.Length, y=Petal.Width, colour=Species)) +
  geom_point() +
  gglayer_theme

The advantage to the custom wrapper approach is the ability to easily mix layers:

gglayer_labs <- list( 
  labs(
    x = "x",
    y = "y"
  )
)

ggplot(iris, aes(x=Petal.Length, y=Petal.Width, colour=Species)) +
  geom_point() +
  gglayer_theme +
  gglayer_labs

Or combine them beforehand:

gglayer_all <- c(gglayer_theme, gglayer_labs)

ggplot(iris, aes(x=Petal.Length, y=Petal.Width, colour=Species)) +
  geom_point() +
  gglayer_all