There is a lot of talk these days about "big" data, but we see a lot of "small data" use cases-- say, 1,000 to 1mm rows-- where repetitive but non-automated tasks hamper everyday work for armies of analysts. Many pieces must be assembled to solve this puzzle, of which this covers a small-- but non-trivial-- bit. Hopefully interesting to at least those lucky folks whose everyday job is to muck around with data and generate these kinds of reports