Take the 2-minute tour ×
Database Administrators Stack Exchange is a question and answer site for database professionals who wish to improve their database skills and learn from others in the community. It's 100% free, no registration required.

I frequently come across this kind of situation where i have a timeseries based data with lots of categories.

What i'd like is some tool that can scan my data and tell me which categories are interesting. Say the measure is a response time for example. If I have 100's of categories, I want to know which ones have the biggest movement in the response time measure.

Now I could do this reasonably easily with Pentaho Data Integration - a tool i'm already a big fan of - but it would be pretty basic. And I can't help but think i'm missing something, and there must be a better tool for the job?

In theory we would then like to do some further advanced analysis with perhaps rstats or weka on the interesting categories - but from my understanding it seems you must identify the interesting categories before using those tools, is that correct?

share|improve this question
    
I think there's a bigger problem: what is interesting? Having values out of an average? Or have some repeated values, showing some pattern ? –  woliveirajr Jun 13 '12 at 12:07
1  
Excellent question. I'd love it if the software could decide itself what is "interesting". Obviously there are a range of ways that could be defined - unusual growth/shrinkage compared to historic change, values out of average, etc etc! Is that asking too much? :) –  Codek Jun 13 '12 at 14:46

Your Answer

 
discard

By posting your answer, you agree to the privacy policy and terms of service.

Browse other questions tagged or ask your own question.