Some analysts have argued that Big Data is fundamentally about data “plumbing”, and not about insights or deriving interesting patterns. It is argued that value (the fifth V) can just as easily be found in “small”, normal, or “weird” datasets (i.e., datasets that would not have been considered before). Do you agree with this? Can you think of small or novel datasets that would provide value as well, without requiring a full-fledged Hadoop set-up?