If she expects at some point to be collaborating directly with data scientists or data analysts in a private-sector setting, then Python is probably the best answer, especially for industries that don’t rely on a doctoral-level talent pipeline from an academic discipline. So, in finance: Python. In healthcare (my sector): Python. In biomedical/bioscience/bioinformatics: probably R and Python.
For both, the current standard data science/data analysis toolsets are add-ons, not the base language. For R, it’s the Tidyverse (way cleaner than base R). For Python, it’s packages like pandas, scikit-learn, and matplotlib. Both languages are open source, which has contributed to their widening adoption over time (sorry, MATLAB; sorry, Stata).
I started using R in 2013. Python was still a year or two away from becoming dominant for data work, and as an ex-academic I looked in academia for data analysis tools. R was easy to stumble across. I’m a product manager, not an engineer or data scientist, so once I learned enough to do exploratory data analysis, data wrangling, and data visualization, I was set. As time went on, I was increasingly aware that Python was taking over that space, but I had what I needed.
I’ve continued to use R because A) switching costs are real, especially since it’s a support tool rather than a thing I spend all day in, and B) the RStudio IDE is soooo great. Along the way, I started reaching for Python when I had clear one-off tasks that were easy to script (like converting the format of video files). I explored Django (a Python web framework) to learn about web applications and be a better partner to my engineering colleagues.
These days, if I had a few weeks of “sabbatical” time, I’d think seriously using the time to reconstitute my R repertoire in Python. That’s because, as others have pointed out, Python itself is a great utility language, although outside of data science it doesn’t have the same mindshare dominance. The R community has developed a lot of “parity” tools that make the ecosystem increasingly more generalist. You can do web scraping in R, you can build websites with R. You can take the scraped data, run a statistical analysis on it, then publish the findings on the website. But there’s a lot to be said for focus when it comes to your tools.