Support other tabular data formats?

[shawnbot]

It would be nice to support reading and writing:

• JSON (array of objects)
• GeoJSON (features in a FeatureCollection)
• Pickled dict lists?
• Protocol buffer repeated lists?

Note that csvkit's in2csv utility already does the first two, so they may actually be less important, and we can point users toward csvkit in the docs.

If we're reading and writing other formats, then we may need to change the utility names (csvorganize -> organ-ize, csvfilter -> organ-filter?).