Good statistical software?

Hello!

I've studied social sciences and newbie in Linux (using Ubuntu in home and windows XP at work), but I was wondering are there any good statistical free/open source software applications?

I downloaded R programming environment but it was too difficult to use... In university we used SPSS, but the license is very expensive and for example in my work I don't need all the functions. (I'm working in a very small company and at the moment using Excel...)

I would be very happy if you have any ideas... please. Excel's driving me crazy...

Hi.

Using keywords statistical analysis package linux spss free for a Google search nets about 50K hits. Among them PSPP, a limited SPSS work-alike, and a list of further pointers to others at Loyola (many of the Loyola links are stale, but have pointers to the right places; not all are Linux-friendly packages).

You probably could recognize keywords better than I can because I don't work in statistics.

The version of JMP from SAS was about $1200, but you usually get some support. With free packages you are on your own, of course, possibly with some help from a forum.

One of the repositories I use (Debian) has a binary instance of PSPP ready to install. Your distribution may or may not.

Best wishes ... cheers, drl

http://www.luc.edu/its/research_software_free.shtml

( edit 1: typos )

You could try RKWard. It's a graphical front-end to the R language that might make it easier to use.

http://linuxappfinder.com/package/rkward

Thank you for your help! I will report later on which software was best for me.

Hi, ubuanu.

After your post, I installed R. Digging through my books, I found that I had a copy of Becker, Chambers, and Wilks, The New S Language. I worked though a few examples to get a feel for it. If you had used SPSS earlier, I can see how you might not like S and R: they are quite interactive. There is very little direct support for report generation, so if that is what you liked about SPSS (in addition to familiarity), then there is little to like about R. However, R is a very *nix-style application in that you could use it as an initial stage from which to generate data for inclusion in reports done by other programs. In the style of *nix, you may need to build some tools of the smaller tools that R provides. The ease of generation of graphics in R is impressive.

I looked at Marsolin's suggestion, RKWard (odd name, that). I can see how it could be considered more friendly, but it appears to be as interactive as R, the difference is that it is a GUI form, not a text form. The little I saw looked a lot like an augmented spreadsheet -- not bad, but not a batching application like the SPSS that I remember.

SPSS was used extensively during the time I was on staff at the University of Minnesota. However, we had statistical consultants to help people with SPSS (and other statistical) questions. The only direct contact I had with SPSS was part of a team that ported SPSS to a Cray-1 (so you can tell how long ago that was!).

Please do post back to tell us of your adventures and final choice(s) in seeking the appropriate statistical tool. Best wishes ... cheers, drl

( edit: clarify )

Hi, ubuanu.

I also found a comparison chart at http://en.wikipedia.org/wiki/Compari...tical_packages which has a few additional pointers. From there I saw that OMNITAB is available from NIST. It is the 80's version, so, as they say on the web page, it does not compare with minitab, which has had continuing development.

My recollection is that OMNITAB is written in Fortran and NIST seems to be making the source available http://www.itl.nist.gov/div898/software/omnitab.html so you could, with some effort use it on your computer.

There are not many free offerings, and many of the packages are for Windows, but if your company had some money, you might be able to obtain something useful. ... cheers, drl

Oh boy, this has been a long journey... (longer than I thought.) Well, I ended up using R programming environment even though I didn't like it at the beginning. Why?

First, it's similar in Linux and Windows environments. This is important since I'm using Windows at work and Ubuntu in home. Secondly, though it took a little time, now I'm quite comfortable with the commands. Big help was that I found a very good R guide in my language. (which is Finnish and the guide can be downloaded from the following address http://users.utu.fi/totavi/) It's a pity I didn't learn all that whilst studying at uni... But better now than not at all.

Good luck for everybody in learning to use R, it's not as hard as you might think :mrgreen:

And thanks for drl! Wikipedia table was very illustrative. There really are not too many free offerings available at the moment... And drl was also right about the easiness of generating graphics in R. In many ways R is actually much better and faster than SPSS. It's also logical once you get started...

Hi, ubuanu.

You're welcome, and thank you for posting back. The description of your journey may well help others. I'm glad it seems to be working out for you.

If you get a chance, post back in a few months on your experiences with R ... cheers, drl

That was a nice post, many thanks ubuana and drl!

I was wondering if there are other R guides available in English with working examples of how to perform specific analyses. I could only spot the sample session in the introductory manual of R at: http://www.r-project.org/ with two working examples. Are there more out there?

Hi.

There is a new O'Reilly book on R: Amazon.com: R in a Nutshell (9780596801700): Joseph Adler: Books ... cheers, drl