Create a function in R to make all the ggplot components more transparent

Question

I am looking for a function that makes my ggplot graphs more transparent. Customizing a new theme I thing that it´s not all good because as the own docs say, 'Use theme() to modify individual components of a theme, allowing you to control the appearance of all non-data components of the plot.'

Supose that we have this simple graph:

Click to copy

ggplot(economics) +
  aes(unemploy, psavert) +
  geom_point() +
  geom_smooth(se = F) +
  ggtitle('Unemploy vs Personal Savings Rate')

It will render like this:

My idea is to build a function that modifies all elements and makes them more transparency with some ratio, something like this:

Click to copy

make.invisible <- function(graph, alpha=.75){
  graph +
    # Change all elements
}

Is it possible with ggplot2?

Answer 1

try this,

Click to copy

print(p, vp=viewport(gp=gpar(alpha=0.3)))

Answer 2

Here is a solution following this post. It uses the edit_colors() function from the colorblindr package. (Disclaimer: I'm an author of that package.)

Click to copy

p <- ggplot(economics) +
  aes(unemploy, psavert) +
  geom_point() +
  geom_smooth(se = F) +
  ggtitle('Unemploy vs Personal Savings Rate')

library(colorblindr) # devtools::install_github("clauswilke/colorblindr")
library(colorspace) # install.packages("colorspace", repos = "http://R-Forge.R-project.org") --- colorblindr requires the development version
# need also install cowplot; current version on CRAN is fine.

# modify alpha values using the alpha function from the scales package
p_alpha <- edit_colors(p, scales::alpha, alpha = .5)

# print
grid::grid.newpage()
grid::grid.draw(p_alpha)

One downside of the alpha() function from scales is that it replaces the alpha value, it doesn't combine alpha values. So here is a simple example of how you can combine alpha values, using your own alpha combination function:

Click to copy

mult_alpha <- function(color, alpha = .5) 
{
  col <- grDevices::col2rgb(color, TRUE)/255
  new_col <- grDevices::rgb(col[1, ], col[2, ], col[3, ], alpha*col[4, ])
  new_col[is.na(color)] <- NA
  new_col
}

p2 <- ggplot(iris, aes(Sepal.Length, fill = Species)) +
  geom_density(alpha = .3) + theme_bw()

p2_mult_alpha <- edit_colors(p2, mult_alpha, alpha = .7)

Comparison between the two approaches shows that they give very different results for plots that already have some transparency:

Click to copy

p2_alpha <- edit_colors(p2, scales::alpha, alpha = .7)

cowplot::plot_grid(p2_alpha, p2_mult_alpha)

(left image: using alpha from scales; right image: using mult_alpha as defined above.)