I have been using Pandas and R for sometime now. I found Pandas is a little rough and fragile around the edges (for example, it was really hard to figure out how to get the median of a distribution in a dataframe as compared to R - fairly basic stuff really). Things do not quite work they way you want them. As for R inspite of the arcane syntax, the quality and number of packages, the help and the IDE (rstudio) is really good.