Different alpha values by facet with facet_wrap in ggplot

Question

I have two graphs from a single long data frame that is faceted based on one of the columns. This produces one facet plot with way way more observations than the other. I want to set a low alpha value for the facet with far more observations but preserve the default alpha for the facet with fewer observations.

Here is a reproduction of my code.

df <- data.frame(spons=sample(c(1:100),10000, replace=T), variable='score', value=runif(10000, -1,1), time=rep(1:10, 1000), DataSub=sample(c('original', 'matched'),10000,replace=T,  prob=c(0.01, 0.99)))
p <- ggplot(df, aes(x=factor(time), y=factor(spons)))
p+geom_point()+facet_grid(DataSub~., ncol=2)

What I want it to look like is the more dense of the plots to have alpha= 0.01 or something like that but leave the more sparse of the plots alone.

Thanks.

Answer 1

You can do this by using the same variable that you used to facet as an alpha aesthetic, then setting the values manually, as follows

p+geom_point(aes(alpha=DataSub))+
   facet_grid(DataSub~.)+scale_alpha_manual(values=c(0.01,1))

This helps, but on my screen at least it's hard to make out much density variation in the upper panel ( values=c(0.02,1) seems a little better: additionally using theme_bw()+ theme(panel.grid.minor=element_blank(),panel.grid.major=element_blank()) helps a bit more by clearing out the background, but you may not want to go that far)