Clickstream

Open Source Statistics to Augment Your BI Tools

I find that periodically I need to do some analysis that requires more stats power than reporting tools like Business Objects or Cognos can provide. When I run into this it either means laborious spreadsheet development or hacking up a one-use program. So today I read up on the R Project for Statistical Computing.

R is an open source stats package that allows you to write statistical processing scripts and the like. You can do more than most BI products allow. In my experience, most BI purchases are aimed at query and reporting, not at the things we do seldom. That makes something like R really attractive.

R looks simple, but the examples I looked at did simple things. Years ago I worked with BMDP, SAS and SPSS to do more complex work. The cost and complexity of these tools is overkill for what many organizations need. I'm going to spend some time with R and see if it's worth adding to the open source BI toolbox to fill in the gaps around the usual BI tool deployments.

There's a good introductory article and mini-tutorial on R by Kevin Farnham at the O'Reilly ONLamp web site to give you a flavor of what R scripts are like and what you can do with it. If you like big lists there's always The Impoverished Social Scientist's Guide to Free Statistical Software and Resources. I like Timothy McSweeney's lists, but they aren't as useful for statisticians.

Posted by Mark Monday, December 05, 2005 9:23:00 PM | permalink |

Comments:

This info is really valuable as my BI experience run into the same problems as you mentioned (using BO or Cognos).

Thanks a lot and please provise more info like this in the future.

# posted by  Anonymous : 9:04 AM

 

Post a Comment