I much prefer R and Julia, but there are algorithms in Stata that don't exist anywhere else. And so, for some jobs, Stata it must be. If you are wondering, for us it was standard errors for clustered panel models.