One of the first suggestions that was presented to me, was by Sean Mulcahy, whom suggested that it would be a good idea to consider confidence regions, as presented HERE (Original Source Not Available Anymore). I am plesed to say that a new geometry (
geom_confidence) has been created, and, is available since version 184.108.40.206, continue reading for a couple of examples.
Here is an ‘out of the box’ example, using Sean’s Feldspar data, for the default 50%, 90% and 95% confidence intervals:
Code for the above:
ggtern(data = Feldspar, aes(An, Ab, Or)) +
geom_confidence() + geom_point() +
labs(title = 'Feldspar - Elkins and Grove 1990 + Confidence Levels')