Using aes_string simplifies the construction of functions to determine the parameters to build:
p <- ggplot(mtcars, aes_string(x="mpg", y="wt", group=interaction("cyl","gear"))) + geom_point()
Now to write a function
make_plot <- function(x,y, interact) { p <- ggplot(mtcars, aes_string(x=x, y=y, group=interact)) + geom_point() }
and call the function
make_plot("mpg","wt",c("cyl","gear"))
But here the interaction is not used, i.e. it is not interpreted. I do not want to use separate variables for bcos interaction, the same can be used for other graphs. How should I make the interaction variable such that ggplot accepts and understands it?
source share