

In [ ]:

    
%matplotlib inline
import datetime
import numpy as np
import pandas as pd
import seaborn as sns

Problem 8.1. Delays at the O'Hare Airpot.

When you are done, rename your file to FirstName_LastName_ord_series.ipynb and submit it via Moodle.
In this problem, you will use Pandas time series functionality to generate sequences of dates from the scheduled departure time at the O'Hare airport and do a time-series analysis of expected arrival delay.



In [ ]:

    
df = pd.read_csv('/data/airline/2001.csv',
                 encoding='latin1',
                 usecols=('Year', 'Month', 'Origin', 'DayofMonth', 'CRSDepTime', 'ArrDelay'))

Function: get_ord_arr_delay()

Write a function named that takes a Pandas DataFrame object, which has Year, Month, DayofMonth, CRSDepTime, Origin, and ArrDelay columns. The function should return a Pandas Series object of ArrDelay at the O'Hare airport (ORD) that has the schedule departure time (Year + Month + DayofMonth + CRSDepTime) as Pandas DateTimeIndex.

To create the DatetimeIndex for Pandas objects, you can either use datetime module or use Pandas to_datetime. The latter link has some useful examples at the bottom of the page.

You should use all of Year, Month, DayofMonth and CRSDepTime columns in the 2001 flights data to create the index.

Don't forget that the ArrDelay column may have some missing values, which you should remove.

In the end, you should get

>>> print(get_ord_arr_delay(df, 'ORD'))
Date
2001-01-01 09:51:00    -9
2001-01-02 09:51:00    -6
2001-01-03 09:51:00    34
2001-01-04 09:51:00    18
2001-01-05 09:51:00   -10
2001-01-06 09:51:00   -14
2001-01-07 09:53:00     1
2001-01-08 09:53:00    -1
2001-01-09 09:53:00   -20
2001-01-10 09:53:00   -17
2001-01-11 09:53:00    -9
2001-01-12 09:53:00   -15
2001-01-13 09:53:00   -16
2001-01-14 09:53:00     4
2001-01-15 09:53:00    -6
...
2001-12-16 16:10:00    23
2001-12-17 16:10:00    78
2001-12-18 16:10:00    -1
2001-12-19 16:10:00     5
2001-12-20 16:10:00    -8
2001-12-21 16:10:00    11
2001-12-22 16:10:00    34
2001-12-23 16:10:00    27
2001-12-25 16:10:00    12
2001-12-26 16:10:00    15
2001-12-27 16:10:00    -6
2001-12-28 16:10:00     0
2001-12-29 16:10:00     7
2001-12-30 16:10:00     5
2001-12-31 16:10:00   -13
Name: ArrDelay, Length: 321784



In [ ]:

    
def get_ord_arr_delay(df, airport):
    '''
    Takes a Pandas DataFrame and an airport
    and returns a Pandas Series of arrival delays.
    
    Parameters
    ----------
    df: A Pandas DataFrame with columns
        'Year', 'Month', 'Origin', 'DayofMonth', 'CRSDepTime', and 'ArrDelay'.
    airport: A str of three-letter airport code.
    
    Returns
    -------
    A pandas.Series object.
    '''
    
    # your code goes here
    
    return result

arr_delay = get_ord_arr_delay(df, 'ORD')

Function: get_daily_delay()

Write a function named get_daily_delay() that takes a Pandas Series, and returns the daily means of the column.

In the end, you should get

Date
2001-01-01    11.929385
2001-01-02    11.277481
2001-01-03    27.546771
2001-01-04     7.348812
2001-01-05     8.911012
2001-01-06    -1.947236
2001-01-07    -1.755149
2001-01-08    -1.424947
2001-01-09    -1.910901
2001-01-10    -2.027311
2001-01-11     0.658281
2001-01-12     7.828633
2001-01-13    -1.032134
2001-01-14    32.323308
2001-01-15    12.141892
...
2001-12-17     4.673966
2001-12-18     2.437198
2001-12-19    14.024125
2001-12-20    10.971084
2001-12-21     2.755030
2001-12-22    17.882812
2001-12-23    14.736909
2001-12-24    -1.488584
2001-12-25    -0.873773
2001-12-26    13.679565
2001-12-27     8.130542
2001-12-28     5.186785
2001-12-29     4.854839
2001-12-30    16.721290
2001-12-31    -0.887160
Freq: D, Name: ArrDelay, Length: 365



In [ ]:

    
def get_daily_delay(ts):
    '''
    Takes a Pandas Series object and returns the daily mean values.
    
    Parameter
    ---------
    ts: A pandas.Series object.
    
    Returns
    -------
    A pandas.Series object.
    '''
    
    # your code goes here
    
    return result

daily = get_daily_delay(arr_delay)

The Plot of Daily Arrival Delay

Use the get_daily_delay function to create a plot of daily mean arrival delays at the O'Hare airport.



In [ ]:

    
# your code goes here

Function: get_hourly_delay()

Write a function named get_hourly_delay() that takes a Pandas Series and a date in string format 'YYYY-MM-DD', and returns the hourly means of the column on the specified date.

In the end, you should get

>>> print(get_hourly_delay(arr_delay, '2001-03-14')
Date
2001-03-14 05:00:00    -6.000000
2001-03-14 06:00:00    -2.553846
2001-03-14 07:00:00     0.756098
2001-03-14 08:00:00    -1.354430
2001-03-14 09:00:00    -3.303571
2001-03-14 10:00:00    -0.155172
2001-03-14 11:00:00    -2.647887
2001-03-14 12:00:00     0.407407
2001-03-14 13:00:00    -1.762500
2001-03-14 14:00:00     1.653846
2001-03-14 15:00:00     2.100000
2001-03-14 16:00:00    10.436364
2001-03-14 17:00:00     9.477273
2001-03-14 18:00:00    14.500000
2001-03-14 19:00:00     3.684211
2001-03-14 20:00:00     1.779412
2001-03-14 21:00:00    -1.941176
2001-03-14 22:00:00    -3.357143
Freq: H, Name: ArrDelay, dtype: float64



In [ ]:

    
def get_hourly_delay(ts, date):
    '''
    Takes a Pandas Series object and a str with 'YYYY-MM-DD' format
    and returns a Pandas Series object of hourly mean values of that day.
    
    Parameter
    ---------
    ts: A pandas.Series object.
    date: A str. In 'YYYY-MM-DD' format.
    
    Returns
    -------
    A pandas.Series.
    '''
    
    # your code goes here
    
    return result

hourly = get_hourly_delay(arr_delay, '2001-03-14')

The Plot of Hourly Arrival Delay

Pick any day you wish, and use the get_hourly_delay function to create a plot of hourly mean arrival delays at the O'Hare airport on that date.



In [ ]:

    
# your code goes here



In [ ]: