Statistical Freeware?

Black Swan

What are the best free statistical software programs available on the web? The university picks up my SPSS license fees, so I was clueless when asked this question at our locale javahouse yesterday.

What are your experiences with statistical freeware? Any pros, cons, limitations, comments, specifications, and links would be helpful.

Squeeonline

If you're using a Mac, you can use graphpad. There's a 30 day free demo, but deleting a certain file resets that clock.

However, to the best of my knowledge no such "work around" exists on windows unfortunately.

Manach

I've used R when I needed to obtain results from statisical data. Free,supported by an engaged community so quite a number of modules and books to support it. A downside is not a user-friendly as some of the commercial packages and there is a learning curve.

actuar90

+1 for R
Takes a while to get into the hang of it but very powerful when you do, I often prefer R to SAS/SPSS

Amirani

Yeah, I was thought R fairly exclusively in college. It takes a while, but eventually you start to love it (sort of).

Ostrom

I'm a stata user with an avowed hatred of R. For an end user like myself, stata's syntax is perfect and I'm happy for my institution to pay the license/upgrade fee every few years for the user support and documentation. However, I've recently started using the QCA module for R and am finding it.... ok. However it's completely un-intuitive to teach beginners with, especially for students who just want to run basic models on secondary data.

Best free program that I regularly use is Gretl, which is brilliant for time series. http://gretl.sourceforge.net/win32/

Black Swan

Has anyone used OpenStat freeware developed by Bill Miller of Iowa State University?

Sarky

R is fiddly and not at all user friendly, but it's free and surprisingly powerful if/when you get the hang of it.

Black Swan

Most everyone seems to be in favour of R stat freeware. Any other stat freeware alternatives that may compete with R? Perhaps be more user-friendly?

Calina

I would favour R as well but it really only fulfills the "free" remit, rather than the "software" remit. It's a programming environment more than user software.

A lot depends on what you want to do. If you're happy to head in the direction of programming, I think the latest version of Python has, or the next version of Python will have some sort of stats module and failing that, you can always look at using SciPy and NumPy. I think Java has some stats libraries as well but I'll drive a long way to avoid using Java so...

If you're focussing on visualisation, arguably, Tableau has a public version but it's web based, and again, on the programming side, D3.js may be an option.

It might also be worth looking at the free edition of Rapidminer although frankly if your dataset is anyway substantial at all, the RAM limit gets annoying very fast. It's more datamining than straight statistical analysis but again, depends on what exactly your user might want to do.

PS: might be worth asking in the BigData/Data Analytics forum as well.

Tree

I do most of my R work in Rstudio, so you might find that takes a bit of the edge off R. There's menus and the like for importing your data.

Squeeonline

has anyone any experience with Sofa Stats? http://www.sofastatistics.com/home.php

It looks quite friendly.

wench

For those who like SPSS, there is a free replacement which may be suitable, called PSPP
http://www.gnu.org/software/pspp/

Black Swan

wench wrote: »

For those who like SPSS, there is a free replacement which may be suitable, called PSPP
http://www.gnu.org/software/pspp/

Have you, or anyone you know, used PSPP? User comments?

Tree

I've used PSPP, I struggled with it as I had never used SPSS either, but the stats guy was able to take to it like a duck to water. I've since learned more about data analysis and like the graphics you get out of R/ggplot2

Black Swan

Has anyone used SurveyMonkey freeware for data collection, and if so, what statistical freeware did you use to analyze the results? Were they user-friendly, or not? Comments please.

dixiefly

This might be of interest.....

http://www.sas.com/en_us/software/university-edition.html

Black Swan

dixiefly wrote: »

This might be of interest.....

http://www.sas.com/en_us/software/university-edition.html

Free SAS? Will investigate this when time permits. Thanks.

Kevster

Regarding R, I'm neither its biggest fan but it's free and you can save your scripts once you have developed them, i.e., for later re-use.

For example, I have a large set of code that I can re-ruse to do factor analysis using K-means and the Gap statistic. I also have code that generates heatmaps and dendrograms, performs microarray analyses (incl. generation of QC images like MA plots and chip images), et cetera.

I am of course happy to share with anyone.

Black Swan

Thanks Kevster. R seems to be one of the most used stat freeware reported. Just wondering if there is something out there more user-friendly not mentioned yet?

Sarky

Tree wrote: »

I do most of my R work in Rstudio, so you might find that takes a bit of the edge off R. There's menus and the like for importing your data.

I tried that out recently myself, it really helps. Much easier to manage directories, plug-in packages, pictures and the like.

Black Swan

RStudio open source statistical freeware appears to meet most of the needs of my colleague with the budget constraints.

James esq

If you want to be an enduser type researcher, survey monkey is a gift for sociology type research, you can get ten multi-questions out to a several hundred users for free and you can analyse up to 100 responses, and a full on survey for for a few euro a week.

For those of us who think statistics is on the way out in certain areas replaced by analysis of large datasets, Rapidminer is yer man. One gig free, very intuitive and similar blackbox approach to GPSS, rated by Gartner, much less expensive than SAS, and doesn't have the steep learning curve R has.

Black Swan

edX offers free online courses of instruction in statistics by several flagship universities. R statistics is a free downloaded programme that we have discussed here in this thread. There is a free introductory course starting 2 Feb 2016: "Foundations of Data Analysis - Part 1: Statistics Using R" offered by University of Texas at Austin that may be of interest.

Link: https://www.edx.org/course/foundations-data-analysis-part-1-utaustinx-ut-7-10x

Manach

I am still a fan of R. Another tool I also would recommend is Elasticsearch/Kibana. It is both a DB and a means to create good user reports from the data. As well, it scales well with large volumes of data. However there is a bit a steep learning curve.

Kurtosis

Black Swan wrote: »

edX offers free online courses of instruction in statistics by several flagship universities. R statistics is a free downloaded programme that we have discussed here in this thread. There is a free introductory course starting 2 Feb 2016: "Foundations of Data Analysis - Part 1: Statistics Using R" offered by University of Texas at Austin that may be of interest.

Link: https://www.edx.org/course/foundations-data-analysis-part-1-utaustinx-ut-7-10x

Thanks for the heads up, I use Stata and SAS (plus a little experience with SPSS) but would be interested in trying out R as I have heard a lot of good things. This looks ideal to gain some familiarity with R to compare with the paid alternatives.

Black Swan

Latest updates on various versions of free R Statistical software for UNIX platforms, Windows and MacOS by the R Project.

tphase

some time ago I had need of something lightweight and scriptable to produce simple statistics on a realtime data stream and came across statist. It's command line only, small (executable is ~174kb) and runs well on low-end hardware which suited my needs for that project.

Black Swan

Free self-paced course "Explore Statistics with R" offered by edX with examples from the health sciences.

paleoperson

I don't get R Studio. I don't have a problem importing data with the command line or clipboard and it's only something you have to do once anyway per project. When I first heard about this software I thought great and hoped for Excel-like functions like data cells editable in real time or quick plots you can build. I prefer the original R look so I went back to that.

Black Swan

paleoperson wrote: »

I prefer the original R look so I went back to that.

Thanks for this observation paleoperson.

Fathom

R Project is free and user friendly.