Off topic - Free Statistics Software

< Next Topic | Back to topic list | Previous Topic >

Posted by Cassius
Dec 6, 2007 at 11:53 PM

I know that there are a few stat practitioners among us. So, if you are interested, http://www.giveawayoftheday.com has Stat Plus for free “today.”

You must download, install and activate the software before 3 AM U.S. Eastern Standard Time.

I have no idea how accurate its calculations are.

-c

Posted by quant
Dec 7, 2007 at 08:49 AM

I’m not much into statistics, but what I hear from colleagues who do, then R is the way to do (also a free sofware), http://en.wikipedia.org/wiki/R_programming_language

Posted by Graham Smith
Dec 7, 2007 at 12:28 PM

quant wrote:
>I’m not much into statistics, but what I hear from colleagues who do, then R is the way to
>do (also a free sofware), http://en.wikipedia.org/wiki/R_programming_language

R is pretty scary for anyone who isn’t a statistician and a programmer. Its command line driven and you need to build up your analysis in stages. So while it is arguably the “best” statistics program available today. ( with some arguments for preferring SAS, Stata or Genstat) it is a very different market from the “easy to use” Excel addin one.

Having said that there are some GUI add-ins for R, but even with these, you quickly need to revert to the command line interface.

Having said that, I would have been interested in trying this out, but time difference means that, as always, I have missed the offer.

Graham

Posted by Cassius
Dec 7, 2007 at 02:57 PM

Graham Smith wrote:
> >R is pretty scary for anyone who isn’t a statistician and a programmer. Its command line driven and you need to build up your analysis in stages. So while it is arguably the “best” statistics program available today. ( with some arguments for preferring SAS, Stata or Genstat) it is a very different market from the “easy to use” Excel addin one.

>Having said that there are some GUI add-ins for R, but even with these, you quickly need to revert to the command line interface.

Graham is correct about R. As I am now officially retired, I have decided not to try to master R…at least for now.

Graham neglected to mention a powerful and intuitively easy-to-use program: SPSS, like SAS, etc. is expensive but pretty easy to use. Several years ago, I had occasion to use it on an out-of-the ordinary problem. Although I hadn’t used it for 15 years, its GUI allowed me to set up 90% of the problem and doing a script for the remaining 10% was quite easy. I haven’t used Stata or Genstat, but I can offer an opinion on SAS:

SAS is a powerful program, but unless it has dramatically changed, its GUI is decades old and it is a PitA (Pain in the A**) to use. Very non-intuitive. And, as someone else has said, the phrase “SAS Help” (referring to its Help file) is an oxymoron. Years after other stat software could easily import Excel files, it offered a 3-day class in how to import Excel data into SAS. Unlike SPSS, one must go into contortions to do anything out of the ordinary in SAS.

-c

Posted by Graham Smith
Dec 7, 2007 at 04:49 PM

Cassius wrote:

While I suspect thee and me are the only one interested in this…

>Graham neglected to mention a powerful and intuitively
>easy-to-use program: SPSS

I have a copy of SPSS on my University PC, but other than a brief trial when I first discovered it was there, I have never felt inspired to use it.

>SAS is a powerful program, but unless it has
>dramatically changed, its GUI is decades old and it is a PitA

I have never used SAS, but as I understand it SAS is still the first choice for really big data sets, Many of the others seem to fall over beyond a certain size. However, this seems to be at least partly a Windows problem, as R in Linux is meant to more capable than WIndows, because of the way Windows manages memory.

Graham

Back to topic list