I have been using R almost daily for years now and its no hyperbole when I say that if ggplot2/dplyr wasn't a thing I would have never bothered with R.
Adding to that, it is what makes me miss R when using other languages. It just isn't very much fun to me to do discovery and building in python, but ultimately where things get operationalized.
I completely agree with you. I think about what I want and R code comes out of my fingers. By comparison, pandas requires more thought to do intermediate-to-advanced things.