Straw Poll on Statistical Computing
The abstract of my previous (reblogged) post claims that R is “the premier language of statistical computing”. That may be true for the wider world of statistics, and I like R very much, but in my experience astronomers and cosmologists are much more likely to do their coding in Python. It’s certainly the case that astronomers and physicists are much more likely to be taught Python than R. There may well even be some oldies out there still using other languages like Fortran, or perhaps relying on books of statistical tables!
Out of interest therefore I’ve decided to run the following totally biased and statistically meaningless poll of my immense readership:
If you choose “something else”, please let me know through the comments box what your alternative is. I can then add additional options.
Follow @telescoper
In the Dark 
December 20, 2016 at 11:28 am
oh well, I am an oldie: f9x/f2k and IDL, lazily finding my way trough python tricks. I know R exists…
December 20, 2016 at 11:33 am
For a number of years I have been using R within Python using the rpy binding (which is out of date now, but still works just fine http://rpy2.bitbucket.org/). It’s possible to use ‘the best of both worlds’ in this case. I like R’s statistical capabilities, but actually use it mostly for making plots, where I think R really excels compared to (e.g.) matplotlib.
December 21, 2016 at 3:34 pm
It was Roald Dahl who wrote James and the Giant Geach…
January 13, 2017 at 3:04 pm
Sorry – missed this. Indeed I am – hope you enjoy!
December 20, 2016 at 11:38 am
Stata for preference. But other options might be SAS, WinBUGS, or JAGS.
December 20, 2016 at 1:35 pm
C++ and ROOT
December 20, 2016 at 1:51 pm
Just switched from IDL to Python. Still making lots of mistakes.
December 20, 2016 at 2:48 pm
C, mostly due to inertia
December 20, 2016 at 5:28 pm
C++/Root
December 21, 2016 at 3:29 pm
This still applies: http://web.mit.edu/humor/Computers/real.programmers
“If you can’t do it in FORTRAN, do it in assembly language. If you can’t do it in assembly language, it isn’t worth doing.”
Seriously, although I do now use python quite a lot, I tend to be nervous of black boxes and for stats problems prefer to code from scratch so I know what’s going on. For me, it will forever be fastest to do that in fortran.
Maybe youngsters will eventually discover “hey, there’s this cool language that’s really very similar to python, except faster”. But then it would become fashionable and in no time we’d have fortran3 that was incompatible with all previous code…
December 21, 2016 at 5:34 pm
Sad old man that I am, I can still remember most of the 6502 Assembly Language Instruction Set, which I learned to program in just after learning basic.