I’d do the pitch correction first. I think when compressing you want to hear the effect of that compression within the context that the audio will ultimately be heard. Compressing after is like compressing what you have, compressing before pitch correction is like compressing what you expect to have later on.
Plus you can change the dynamics in Melodyne in addition to the pitch which is basically like being able to compress individual blobs as desired. If you edit the dynamics in Melodyne you can tame any blobs that are too loud (or soft). Then if you later apply compression it won’t be to fix outliers but rather to congeal the entire vocal.