Create a separate time series command for aggregating just over time

The talk by Cristoph Paulik dealt a lot with this ( http://splendidabacus.com/2015-eoscience2.0/). He also talked about the problem of storing ungridded time series data. Talked about using contiguous ragged (or indexed ragged) netCDF (CF compliant) data sets. This is all done by a package called pygeobase. It also does some clever temporal traversal of data given a spatial region.

This should be easy with pandas though...