Just wanted to follow up with an example:
df = data.frame(y = c(rnorm(200)*30, rnorm(100)), x =
factor(c(rep(c("class1","class2","class3","class4","class5"),length.out=200),rep("class6",100))))
ggplot(data=df, aes(x = x, y = y)) + geom_violin()
It seems that geom_violin scales the areas of each shape such that
the width of the widest violin is a constant. In the case above,
this scaling makes the thinner violins difficult to interpret. It
would be nice to be able to modify that maximum width such that the
total areas of each shape is increased, and all shapes become a bit
fatter as a result.
Any thoughts?
Thanks,
allie