Replicating pivot_table with a groupby aggregation

At first glance, it may seem that the .pivot_table method provides a unique way to analyze data. However, after a little massaging, it is possible to replicate its functionality with the .groupby method. Knowing this equivalence can help shrink the universe of pandas functionality.

In this recipe, we use the flights dataset to create a pivot table and then recreate it using the .groupby method.

How to do it…

1. Read in the flights dataset, and use the .pivot_table method to find the total number of canceled flights per origin airport for each airline:

   >>> flights = pd.read_csv('data/flights.csv')
   >>> fpt = flights.pivot_table(index='AIRLINE',
   ...     columns='ORG_AIR',
   ...     values='CANCELLED',
   ...     aggfunc='sum',
   ...     fill_value=0)
   >>> fpt
   ORG_AIR  ATL  DEN  DFW  IAH  LAS  LAX  MSP  ORD  PHX  SFO
   AIRLINE
   AA   ...