Create Sample Time Series Data in Python (DateTime)

I’ve been playing around with time series data in Python and was looking for an easy and clean way to create random samples of datetime objects.

There’s more than one way to do this, but I was particularly fond of an approach using the faker library, which is outlined on StackOverflow here.

Code Example

First, we’ll import the faker library and instantiate an Faker object to generate random samples.

import faker
fake = Faker()

Then, we can test out creating a sample datetime like so:

fake.date_time_between(start_date='-10y', end_date='now')
# datetime.datetime(2014, 10, 1, 6, 58, 47)

Or, if we want just a date object, we can do that too:

fake.date_between(start_date='today', end_date='+10y')
# datetime.date(2023, 2, 22)

Now, if we want a range of samples, we can use Python’s array creation syntax:

datetimes = [fake.date_time_between(start_date='-30y', end_date='now') for i in range(5)]
datetimes

# Output:
# [datetime.datetime(2008, 12, 2, 15, 0, 5),
#  datetime.datetime(2006, 4, 14, 12, 44, 6),
#  datetime.datetime(2019, 10, 1, 17, 45, 42),
#  datetime.datetime(2009, 9, 25, 15, 22, 49),
#  datetime.datetime(2001, 3, 6, 8, 10, 3)]