Use dplyr approach to summarise()

[MichaelChirico]

Follow-up to #520 to remove another {plyr} import (plyr::here). Note that this branch includes #520 -- GitHub doesn't allow a way for me to stack PRs from my fork.

Tested equivalency locally for this example:

ggally/man/ggally_summarise_by.Rd

Lines 49 to 50 in 9d954c1

	data(tips)
	p_(ggally_summarise_by(tips, mapping = aes(x = total_bill, y = day)))

And also by (1) adding a row with missing y (2) adding a level to the y factor that's not present in the data.