Locally, we continue to use SAS for manipulation of linked health and population registers. The combination of the SAS data step, SQL and macros is powerful for more complex data. However, most of our researchers use R or Stata for data analysis.
That said, younger researchers often prefer to use dplyr or data.table in R for the data management. One challenge is that those researchers often can not read the legacy SAS code - hence they do not understand how to handle the corner cases. This may be a symptom that data management documentation is often poor in academia.
For data analysis of time-to-event data, Python and Julia both lag behind R, Stata and, to an extent, SAS. I would be happy to move to Julia, but some of the basic modelling tools (e.g. safe prediction for formulae with splines) are currently lacking.