Best Statistical / Graphing Software Package

Minitab

SAS

I really like STATA, simple to use and powerful but SAS seems to be preferred by more, even though it is harder to learn in my opinion.

SAS JMP may be worth a look.

In my experience, the major stat programs will do 90% of what you want to do well. There will always be some things that it just won't do or does it in the most ridiculous and obtuse manner that it just isn't worth it. You may have to resign yourself to getting more than one program.

My work computer currently has:

SAS: does everything; expensive; bitch to work with

SAS JMP: convenient; easy to use; can be tough to integrate output with other programs like Powerpoint

SPlus: very good for graphing; rarely use the analysis portion

SpotFire: does some amazing graphing; no analysis portion

MiniTab: easy to use; prefer to use JMP though

SPSS: rarely use

Statistica: rarely use

Faster Analytics: used occasionally for Bayesian nets

XLStat: handy for some types of analysis

Latent Gold: use occasionally for latent class analysis

I'm glad that I didn't have to pay for all of them.

I love stata. Seems like it's becoming the go to package in social sciences. If it does everything you need, then go for it. Sas does everything, but its expensive and harder to learn.

I'd consider going Stata for its affordability (180 if you're a student and don't need to work on giant data sets). it's really a great program. Its usercommunity is vibrant and it has loads of free addins. It's lightweight, too.

SAS for stats and sigma plot for graphs

Thanks for some feedback.

Have any of you used R? It seems to me that R can do ANYTHING, including stats and graphs. The problem is, like SAS, it has a huge learning curve. I currently use R to do all of my linear mixed effects models. I am even planning on using R to do latent class; it is apparently as good as Latent Gold for this.

I like SPSS because it is point and click. What requires lots of figuring out how to do through code in R, I can do in SPSS in a few seconds. But I have found SigmaPlot to make nicer graph than both of these.

It sounds like there is no one perfect program that does it all. I guess my goal would be to find the best graphing program then, and it sounds like SigmaPlot is a front runner.

If you want to make nicer graphs in R with less effort try the ggplot2 package. http://had.co.nz/ggplot2/http://had.co.nz/ggplot2/book/qplot.pdf

If you are going to continue to use statistics then take the time to learn R. Its free and can do anything. Build up some graphic template functions. Jmp by the way is very tedious to manipulate plots but is otherwise a fun little program.

Soon I'll be taking a Business Analysis course, which is mostly stats, for my MBA degree, and I am wondering if anyone has an opinion about which stats software to use. I'll have to write essays, including stats results/analysis, so I may need a good graphing software package. Any advice is greatly appreciated.

Thank you,

Tinman

runfastcoach@gmail.com

Rather than working in several packages, just learn R. The learning curve is not as bad as people make it out to be. Once you learn the basics, you can get code for almost anything. Modifying existing code is usually straightforward if you know what you're doing statistically.

I used SAS, Statgraphics and Statistica over the last 20 years, but once R got rolling I switched. I think it was well worth the effort. The flexibility is great, and most cutting edge stuff (in my field, anyway) is available in R before the commercial packages.

Any other recommendations? It sounds like I should put some time into R, and maybe still consider SigmaPlot for graphics.

I have tested many statistics and graphics software packages. GraphPad Prism 6 is the most intuitive, handles most routine statistics, produces publication-quality 2D graphs, and runs on both Mac and Windows. OriginPro 9 is less intuitive than GraphPad Prism 6, but it has more statistics and graphing features; however, it runs only on Windows. Statistica 10 has the most comprehensive set of statistical functions and graphing options, although it does not offer quite the control over graphics provided by GraphPad Prism 6 or OriginPro 9; Statistica 10 runs only on Windows.

R is by far the best statistical programming software. I am quite perplexed by the responses here. There is a learning curve, for sure, but it is the most powerful by far and it is the most used by actual statisticians, researchers, and academia. It is probably the most used in finance/alpha-generation as well, which is where I would look to find what is the best available.

i second graphpad prism for publication-quality graphs/plots.

SAS is the best. You must be a flunky if you have to as this question.

Lotus 123

If you want to stay current, R is clearly the way to go these days. The advantages of avoiding a menu-driven program vastly outweigh the learning cost, which is not early as bad as people make it out to be. A year ago I just didn't install any of my old stats packages when I got a new computer. With that incentive, I became proficient in about a month.

Install R studio and learn to use it there, rather than the plain R gui.

http://www.rstudio.com/

(watch the 5 minute tutorial, and you'll be ready to get rolling)

http://www.springer.com/statistics/computational+statistics/book/978-0-387-93836-3

(free, basic intro if the use-r books are available to your institution).

Some good links that helped me get rolling in R:

http://www.cookbook-r.com/

(good beginner's guide with example code: cut and paste code into the R studio editor and execute the code one line at a time, and you'll soon see what's going on.)

http://www.ats.ucla.edu/stat/dae/

(very useful for code you can modify. Also basically a good statistical reference.)

http://stackoverflow.com/questions/tagged/r

(use the search function with this one, for very specific questions)

http://www.unil.ch/webdav/site/ee/shared/WorkshopD.pdf

(good tutorial on mixed models in R by Bates).

Also, the post above about ggplot2 for graphics within R is right on. Far more flexible and comprehensive than any commercial package I've used over the years. The easiest way to learn ggplot2 that I've seen is through the 'cookbook for r' link in the last post... otherwise the documentation is a bit opaque at first. Once you've built up a set of scripts that do the tasks that you do often, modifying them is fast and easy. If you google '_____ ggplot2 r' you'll usually find code that will do what you want.

If you're going to use R, you should use R Studio. Much better interface.