I my opinion, a big chunk of "data science" work is also the collection of data and eventually some sort of deployment. R is great, but in my opinion, it is functionally a little bit limited to the middle part: The statistical analyses and predictive modeling/machine learning.
If you need to use/build a data scraper or want think about developing e.g., an online app for your classifier, you are way better off with Python. I am saying this as a former R guy who made the switch a couple of years ago...
After finishing 3 Course courses of R I have decided that I dont want really work in R and I switched to Python. Its much more intuitive than R, its real programming language, its modern.
Sure. But the point of the article is that R is the de facto language for data science and is strongest than the alternatives in key areas such as data manipulation, data visualisation and machine learning.
Hmm. Parent's point is still valid. A language can be (a) excellent at what it was designed to do and (b) poor at what it wasn't designed to do. R is a statistical language first and foremost. My experience is that R is not used in production systems, and has serious PL warts compared to other languages.
Also note: between scikit-learn and pandas, Python is also excellent at ML and data manipulation.
A) Build on what you know. If you have done any programming before, Python is probably the way to go. Otherwise R is probably fine.
B) Learn enough of the other language to at least read it.
C) Always be learning and willing to learn.
In my (anecdotal) experience, Python data skills seem to be in great demand. I don't remember talking to anyone in my most recent job search that was looking for only R. Typically it was Python or Python and R. This may be colored by the fact that my experience is mostly in the San Francisco tech scene.
My observation is that use of R vs Python seems to hinge a lot on people's backgrounds. People with backgrounds in engineering and the physical sciences seem more likely to use Python. People with backgrounds in stats, medical science, or econ are more likely to go with R.
Do you think Python isn't popular in academia? I'm not sure I agree--I saw lots of Python written by faculty, TAs and researchers during my MS. Not trying to say it's more popular than R/MATLAB, but I think Python is "sufficiently" popular in academia at this point.
I think this depends very much on what area of academia.
I had heard of R from reading online, but in my time as a physicist I never came across anyone using it. Most of the discussion was around Python vs Matlab, though you also dealt a lot with C vs C++ vs Fortran... :)
