First of all, the question is 'Why Use R'. One source answers that question thus:
R is the leading tool for statistics, data analysis, and machine learning. It is more than a statistical package; it's a programming language, so you can create your own objects, functions, and packages.
Speaking of packages, there are over 2,000 cutting-edge, user-contributed packages available on CRAN (not to mention Bioconductor and Omegahat). Many packages are submitted by prominent members of their respective fields.
[More....]
For Beginners in R, here is a 15 page example based tutorial that covers the basics of R.
These tutorials are designed for beginners in R, but they can also be used by experienced programmers as a refresher course or as reference. Running loops in R can be slow and therefore the apply group of functions as well as the reshape package can drastically improve the performance of the code.
(Score: 5, Interesting) by physicsmajor on Saturday August 27 2016, @10:57PM
Don't get me wrong, R is a fantastic tool. However, their claims are more than a little overblown.
R is probably the leading tool for statistics. No argument there.
Data analysis, though? Strongly disagree. Python is equivalent or superior to R from the standpoint of data structures (Pandas DataFrames for those who like R, NumPy arrays for those who don't, Dask variants of both of these for truly Big Bata, XRay for those who like N-dimensional typed data) which are essentially equivalent or more general and easier to handle than R. The surrounding general purpose scientific tools in the ecosystem are also far ahead on the Python side.
Machine learning? No. Just no. For classical machine learning there isn't any coherent argument that R is superior to Python with scikit-learn; everyone is using scikit-learn today. For Deep Learning, I don't think R even has a framework to interface with TensorFlow. The only one I could find went through Python.
I do get that the author is excited about R, and wants to evangelize, and I encourage this. R is a great tool! But it isn't the perfect general purpose science tool that opening headline might make someone think it is. R is a fantastic statistics package with some general science added on. For general science, Python is just better (and can directly interface with R for esoteric statistics when needed). Maybe I'm being overly pedantic, but I do think we should demand honesty up front.
(Score: 0) by Anonymous Coward on Sunday August 28 2016, @12:33AM
> R is probably the leading tool for statistics.
Would you say it's more popular and/or better for statistics than MATLAB?
> Machine learning? No.
I read on the Web that GNU Octave is better for machine learning than R. Do you agree?
(Score: 3, Informative) by physicsmajor on Sunday August 28 2016, @12:39AM
Matlab is junk for statistics. Absolutely terrible, you might as well be using Excel. Nothing even approaching a DataFrame; R is light-years ahead of Matlab. IGOR Pro is probably the best (though not all that well known) commercial alternative to R for DataFrame-like analysis. For pure stats think JMP or SAS which are full-featured but a pain in the ass to use compared to R.
Not sure about Octave.
(Score: 0) by Anonymous Coward on Sunday August 28 2016, @08:31AM
Octave is a clone of Matlab. It's selling point is that it is completely free. So if Matlab sucks, then Octave will suck.
(Score: 1) by kanweg on Sunday August 28 2016, @10:24AM
I used IGOR many years ago and it was really great. I can only assume that it has not gotten worse.
Fun thing (don't know if it is still in), you could change the history of what you'd done. To do confirm, you got a dialog box saying: Do you want to change history, with to buttons: Da and Njet.
Bert
(Score: 3, Informative) by cellocgw on Sunday August 28 2016, @04:40PM
> R is probably the leading tool for statistics.
Would you say it's more popular and/or better for statistics than MATLAB?
Having used both R and MATLAB quite extensively, I can tell you that R wins hands-down. Here are some of my top reasons (leaving aside the absurd cost of MATLAB)
-- R's function argument syntax is miles better, more flexible, and more intuitive than MATLAB's .
-- R understands NAMESPACEs and ENVIRONMENTS; Matlab is clueless.
-- To load functions in R , one sources them. Everything in a sourced file is available at the command line. None of this "only the top function is available" crap as in MATLAB.
-- ggplot . 'nuff said
-- Perfectly legal in R: x = func(y,z)(a,b)[[3]][5:9] . That is, func returns a function, which acts on a&b, returning a structure (list) of which we took the 3rd elemnt and took the 5:9'th elements of it. Matlab won't even let you do x = sin(y)(3:5) .
-- R allows you to access and modify portions of functions (actually, closures) in real-time, sort of like Mathematica. MATLAB doesn't even let you define a function from the command line.
R's source code is openly available. Neither MATLAB's engine nor the "builtin" functions' codes are available.
Physicist, cellist, former OTTer (1190) resume: https://app.box.com/witthoftresume