Add discussion of pandas (in lieu of numpy?)

[ariddell]

Pandas does make many operations much easier. Need to find sensible ways of integrating mentions of its uses. In principle, I think the tutorials should only require familiarity with the "basic" numpy/scipy stack.

[ariddell]

Anyone who is interested in this tutorial is probably ready for pandas. Upon reflection, I don't think it adds too much complexity.

[christofs]

I also think including pandas is a good idea. DataFrames and label-based slicing of them is very useful in our context and actually makes things a lot more intuitive.

[ariddell]

How should I weave it in? Should there be a separate tutorial showing slicing by label etc?

[christofs]

Not sure about this. Maybe a brief section on dealing with tabular data being output by Mallet or NMF could be added somewhere and then referenced in the various relevant places: reading such data into a pandas dataFrame, slicing by label, etc. It is kind of a bridge between TM/NMF and the visualization part, so maybe it could fit at the beginning of the Visualization chapter, as well.