The R language isn't for the casual end user. Data scientists write in R to perform complex statistical data analysis. So how do you make it enterprise-friendly? David Smith, the vice president of marketing for Revolution Analytics, explains to IT Business Edge's Loraine Lawson how the company's Nov. 15 release of Revolution R Enterprise 5.0 makes Big Data analysis more accessible. Smith also writes the company's blog, Revolutions.
"Our deployment strategy is around empowering those data scientists using R to create little apps that can then be embedded into, say, a BI dashboard or Microsoft Excel. We do that by creating a server version of Revolution R Enterprise with a Web services API."
Lawson: You're presenting your solution as part of Big Data stack. What's that mean?
Smith: We're really positioning Revolution R Enterprise as that analytic layer within the stack. We're agnostic as to where data is actually stored, but we've been building a lot of connections to various places where people might be storing data, especially Big Data.
We announced the connection with the IBM Netezza devices a couple of months ago now. Something that we brought out recently was the integration with Hadoop and, in fact, we've certified that with the Cloudera CDH3 distribution that we supported with our 5.0 release. We have also announced integration with the Microsoft HPC server platform for doing high-performance, distributed computing, which is also part of this release.
Lawson: Your solution is an analytics tool. Do the people who use it need to know how to write R?
Smith: Yes, they absolutely do. The typical user of the Revolutions enterprise software is a statistician, a data analyst or a quantitative analyst or, a more modern term for all of those things is a data scientist - somebody who has the expertise in doing predictive modeling and has an understanding of the data they're working with and the computer science skills to actually get access to that data and prepare it for analysis.