Yeah, totally. I spend a lot more of my time writing Python for anything that isn't data exploration/analysis, for exactly that reason.

I still refuse to learn pandas enough to replace dplyr though, as it's just so painful to use the API compared to how easy this stuff is in R.