Norman Nie, co-creator of SPSS (acquired by IBM for $1.2 billion last summer), and his group Revolution Analytics aim to bring analysis to a wider audience with a product built on top of R, the popular statistical computing language. They call it Revolution R.
Noted in a recent Forbes article:
R is a powerful tool but difficult for novices to use. Nie’s Revolution Analytics aims to make it more accessible with a better-organized library, capabilities for bigger jobs and a user interface that lets users drag and drop statistical analyses into place, outputting easily read charts.
The rest of the article is about Nie, the growing importance of data, etc.
I’m curious. Has anyone tried Revolution R? They say that it has “faster performance and greater stability” than base R. Is it that much better?
[Thanks, Victoria]
I used it briefly. For some reason I wasn’t able to simply install the RMySQL package on it, so I’ve scrapped it for now. Nothing wrong with trusty ol’, regular R. I imagine they are doing some pretty cool stuff though. I’m definitely going to keep an eye on what they have to offer.
They are doing some interesting stuff to make R work ‘painlessly’ on clusters and the like– and split large calculations across nodes, and apparently improved Math libraries. Really stuff you wont notice unless you are using ‘BIG’ data
On some platforms they have a nicer IDE and some web integration tools for R.
Oh and for an academic the full version is free (so I was planning on giving it a go…. when I fix our server).
Regarding… things not working (e.g. RMySQL).. it’s always been my experience with bleeding edge R that glitches happen and get sorted after a bit. I always install a new version and keep the old for a few weeks till fixes trickle through the community. Best do the same with Revolution.
I installed it since there is a free license for academic users. Unfortunately it only runs on Windows and Redhat…
Revolution R is more of an IDE for R than anything else. Given how it is billed in the Forbes article was was expected a better version of Rcmdr or something that would be familiar to SPSS users.
I work for Revolution Analytics, editing the Revolutions blog. Thanks for the mention, Nathan.
The faster performance of Revolution R comes from the fact that Revolution R is compiled to link with the multithreaded Intel MKL libraries. Basically, this means that for linear albegra and matrix functions (which is what many of the stats functions in R boil down to), R will use optimized algorithms *and* use all available processors if you have more than one. You can see some the comparisons with regular R (which ordinarily only uses one processor on Windows) in these benchmarks.
I’d also mention that we have big plans to enhanced Revolution R further, with more support for really large data sets and an easy-to-use GUI. We’re released some of the details in this paper.
Finally, as mentioned above, if you’re in academia you can download our full commercial edition, Revolution R Enterprise free here.
Dan, your comment prompted me to publish this post on how to connect Revolution R with MySQL.
Sweet! We can only hope that the UI is as sexy as SPSS.
The one thing I like about R is that it strips away all the junk native in SPSS. Software that focused on UI could dominate this space. SAS, SPSS and any/all other alternatives would go straight down the drain. Give a cohort of grad students (5-10 years) a chance to use it and its game over.
So they want to make a free SPSS-like version of R now? Well, I am pretty sure they will charge for it. Of course R is not as easy to use as SPSS but it is so flexible and for free which makes it great.