One plus with dplyr is that I can share the code with non-R programmers (and even some non-programmers) and they can follow what is happening pretty easily, while data.table takes some more explanation.