On Fri, 26 Mar 2010, Florin Iucha wrote:

> On Fri, Mar 26, 2010 at 12:15:40PM -0500, Mike Miller wrote:
>
>> It always depends on what your goals are.  For aspiring scientists, 
>> especially those who need to do statistical and/or probabilistic work, 
>> I recommend learning GNU R.  It also makes nice graphs.  It is the king 
>> of the stat packages now and that will continue.
>
> Matplotlib (a python library) by default produces better-looking graphs 
> than R.  You can of course tweak both, but for quick graphs I prefer the 
> former.

Very interesting.  Are there good screenshots somewhere?  Is it hard to 
learn to make graphs in Matplotlib starting from, say, a tab-delimited 
data file?


>> Once they've learned some R, if they want more flexibility for more 
>> general kinds of programming, I would recommend Python.  I'm not using 
>> it myself, but it seems to be the way to go today.  I've talked with a 
>> lot of people about it and that's where I stand.
>
> Even if you mostly use R for stats, you still need a good bit of python 
> to munge the data files - R is nice for importing CSVs, but lots of 
> manipulations are painful.

True, but it depends on the original formt of the data.  I usually use 
perl because I don't know python.  R has a lot of import functionality in 
the base package and in add-ons.

Mike