featuretools.demo.load_flight(month_filter=None, categorical_filter=None, nrows=None, demo=True, return_single_table=False, verbose=False)

Download, clean, and filter flight data from 2017. The original dataset can be found here.

  • month_filter (list[int]) – Only use data from these months (example is [1, 2]). To skip, set to None.

  • categorical_filter (dict[str->str]) – Use only specified categorical values. Example is {'dest_city': ['Boston, MA'], 'origin_city': ['Boston, MA']} which returns all flights in OR out of Boston. To skip, set to None.

  • nrows (int) – Passed to nrows in pd.read_csv. Used before filtering.

  • demo (bool) – Use only two months of data. If False, use the whole year.

  • return_single_table (bool) – Exit the function early and return a dataframe.

  • verbose (bool) – Show a progress bar while loading the data.


In [1]: import featuretools as ft

In [2]: es = ft.demo.load_flight(verbose=True,
   ...:                          month_filter=[1],
   ...:                          categorical_filter={'origin_city':['Boston, MA']})
100%|xxxxxxxxxxxxxxxxxxxxxxxxx| 100/100 [01:16<00:00,  1.31it/s]

In [3]: es
Entityset: Flight Data
    airports [Rows: 55, Columns: 3]
    flights [Rows: 613, Columns: 9]
    trip_logs [Rows: 9456, Columns: 22]
    airlines [Rows: 10, Columns: 1]
    trip_logs.flight_id -> flights.flight_id
    flights.carrier -> airlines.carrier
    flights.dest -> airports.dest