R - add transparency to colorRampPalette

Question

I'm using smoothScatter to plot some data. I'm trying to change the color of the density from the default of colorRampPalette(c("white", blues9) to transparent red. How can I do this?

I tried:

smoothScatter(
  x,y,nrpoints=length(df$x),
  colramp = colorRampPalette(c(rgb(1,1,1,0.5),rgb(1,0,0,0.5))
)

But this doens't work.

More generally put, how can I create a colorRampPalette function that can be passed in as a function that includes making all the colors transparent?

Answer 1

You just need to add alpha=TRUE to your call to colorRampPalette:

>?colorRampPalette
   alpha: logical: should alpha channel (opacity) values should be
          returned?  It is an error to give a true value if ‘space’ is
          specified.

> colorRampPalette(c(rgb(1,1,1,0.5),rgb(1,0,0,0.5)))(3)
[1] "#FFFFFF" "#FF7F7F" "#FF0000"
> colorRampPalette(c(rgb(1,1,1,0.5),rgb(1,0,0,0.5)), alpha=TRUE)(3)
[1] "#FFFFFF80" "#FF7F7F80" "#FF000080"

Since your colorRampPalette doesn't seem to have the alpha argument, here's a manual solution:

# Specify alpha as a percentage:
colorRampAlpha <- function(..., n, alpha) {
   colors <- colorRampPalette(...)(n)
   paste(colors, sprintf("%x", ceiling(255*alpha)), sep="")
}
colorRampAlpha(c(rgb(1,1,1),rgb(1,0,0)), n=3, alpha=0.5)
[1] "#FFFFFF80" "#FF7F7F80" "#FF000080"

You'll need to specify the number of colors you want in advance however.

Answer 2

Here is my inelegant way to create a transparent palette with colorRampPalette

## first specify your color scheme

colorRampPalette(c("lightgrey","grey","lightblue","blue","darkblue"))  -> blues

## Then create the palette as you would normally
blues_pal<-blues(500)

## Finally use paste to set transparency to every color in the vector
blues_pal_transp<-paste(blues_pal, "50", sep="")

## and an example
image(volcano, col=blues_pal_transp)