Applied dimensionality

Acquisition trend affects open-source world as well -- Pentaho "acquired" Weka, data-mining toolset to put another brick in their all-covering-BI-wall.
Weka is a new name to me, so I'll toy with it for a while. Just need a sample dataset to get started. Only one at hands is aggregated list of client site EP errors (the so-famous PlanningErrorLog.csv). Around 100 mbs of raw data, nearly 250k of records -- maybe there is something we can't figure without using data mining. Except, of course, the fact that it Cognos EP is buggy.