Megabus famously advertises $1 fares on the sides of their buses. If you’ve ever driven on I95, you’ve probably seen them. This jupyter notebook fetches data from their website in search of these elusive fares.

megabus credit: megabus

origin = 'DC'
destination = 'NYC'

#papermill_description=define_functions
import pandas
import requests
import datetime

def get_id(name):
    d_name_to_id = { 'NYC': 123,       
        'Philadelphia': 127,      
        'Baltimore':   143,
        'VA Beach': 475,
        'DC': 142 }

    return d_name_to_id[name]

def compute_date(delta=0):
    """return date 'delta' days from now in format %Y-%m-%d, e.g., 2022-12-02"""
    now = datetime.datetime.now()
    dt = now + datetime.timedelta(days=delta)
    return dt.strftime('%Y-%m-%d')

def get_data(departure_date, origin, destination):

    origin_id = get_id(origin)
    destination_id = get_id(destination)

    """returns a dataframe from megabus with the rides on the indicated departure_date"""
    # need to override default requests header otherwise Megabus will categorize the request as a bot
    headers = { 'User-Agent': 'Mozilla/5.1'}
    url_path = 'https://us.megabus.com/journey-planner/api/journeys'
    r = requests.get( f'{url_path}?originId={origin_id}&destinationId={destination_id}&departureDate={departure_date}&totalPassengers=1&concessionCount=0&nusCount=0&otherDisabilityCount=0&wheelchairSeated=0&pcaCount=0&days=1')

    # fix dataframe
    df = pandas.DataFrame(r.json()['journeys'])
    df.set_index('journeyId', inplace=True)
    df.drop(['legs', 'origin', 'destination', 'reservableType', 'lowStockCount', 'routeName',  'duration', 'serviceInformation', 'promotionCodeStatus'], axis=1, inplace=True)
    df['origin'] = origin
    df['destination'] = destination

    df.departureDateTime = pandas.to_datetime(df.departureDateTime)
    df.arrivalDateTime = pandas.to_datetime(df.arrivalDateTime)

    df.price = df.price.astype(str).astype(float)

    # re-order columns
    df = df[['origin', 'destination', 'departureDateTime', 'arrivalDateTime', 'price']]

    return df

# fetch round trip data

df = pandas.DataFrame()


for i in range(0, 65):
    try:
        df = pandas.concat([df, get_data(compute_date(i), origin, destination)], axis=0)
        df = pandas.concat([df, get_data(compute_date(i), destination, origin)], axis=0)
    except:
        break

max_day = i
# df

Fare Analysis

print(f'Analyzing fares between {origin} and {destination} for the next {max_day} days')

Analyzing fares between DC and NYC for the next 45 days

Fares < $10

df[df.price < 10]
origin destination departureDateTime arrivalDateTime price
journeyId
19903121 NYC DC 2022-12-31 18:00:00 2022-12-31 22:30:00 1.0
19903505 NYC DC 2022-12-31 21:30:00 2023-01-01 01:50:00 1.0
19899024 NYC DC 2023-01-09 06:00:00 2023-01-09 10:20:00 1.0

Fares Distribution

import matplotlib.pyplot as plt

df.hist(column='price', by='origin', width=5, grid=True, sharex= True, sharey=True, density=True, figsize=(20, 5));
plt.suptitle('Density Plots of Prices from Origin City');

png

Jupyer Notebook Automation/Conversion Notes

papermill can run this notebook in batch-mode,

$ papermill megabus.ipynb megabus_out.ipynb -p origin DC -p destination DC

To convert this file to markdown without the Python code,

$ jupyter nbconvert megabus_out.ipynb --no-input --to html