Creating Fake (Mock) Data with Python

In this post I wanted to share an interesting Python package and some examples I found while helping a client build a prototype. The package generates mock or fake data and is simple to use and works well if you need to quickly generate some dummy or mock data for testing purposes.

From my perspective, it can also be useful for teaching analytics and also if you want to generate fake or mock up data with many different data types easily and quickly. You can then export it as a .csv or use pandas data frames for other data science and analytics use cases.

It contains some great documentation too: Faker’s documentation

A few examples I wanted to share are shown below. Start by importing the Faker library and pandas:

from faker import Faker
import pandas as pd

Here we initialise Faker generator and create an example of generating a fake data for a random name:

faker = Faker()
faker.name()

'Eric Poole'

You’d probably want to generate more than one fake data record at a time:

for n in range(5):
    print(faker.name())

Whitney Davies
Christopher Johnson
Robert Washington
Zachary Williams
Mark Ramirez

There are many other different types of fake data you can generate by using ‘providers’. For a list of providers see the docs.

from faker.providers import internet, geo

You can generate fake mock data on latitude and longitude, for example:

for n in range(5):
    print(faker.local_latlng(country_code="AU", coords_only=False))

('-31.95224', '115.8614', 'Perth', 'AU', 'Australia/Perth')
('-37.88333', '145.06667', 'Carnegie', 'AU', 'Australia/Melbourne')
('-32.05251', '115.88782', 'Willetton', 'AU', 'Australia/Perth')
('-32.05251', '115.88782', 'Willetton', 'AU', 'Australia/Perth')
('-33.75881', '150.99292', 'Baulkham Hills', 'AU', 'Australia/Sydney')

And if you wanted to generate a custom list of fake data types and create a pandas data frame with these fake data points:

df = []

for n in range(5):
    df.append({'Lat': faker.coordinate(center=74.0, radius=0.10),
               'Lon': faker.coordinate(center=40.8, radius=0.10),
               'Txt': faker.sentence(),
               'Nam': faker.name(),
               'Add': faker.address(),
               'Job': faker.job()
              })

df = pd.DataFrame(df)
df = df[['Nam', 'Job', 'Txt', 'Add', 'Lat', 'Lon']]
df

There is also an option to bring fake data about a profile pre-built by the package:

faker.profile()

{'job': 'Architect',
 'company': 'Martinez, Cruz and West',
 'ssn': '311-64-2980',
 'residence': '3864 Sanford Dam Suite 803\nWest Leroy, MS 46211',
 'current_location': (Decimal('-38.669420'), Decimal('86.843925')),
 'blood_group': 'B+',
 'website': ['http://church.com/',
  'https://www.mayer-maldonado.com/',
  'http://matthews.org/'],
 'username': 'ocannon',
 'name': 'David Johnson',
 'sex': 'M',
 'address': '1431 Berry Extensions\nSouth Jameshaven, ME 13555',
 'mail': 'wilkinssteven@gmail.com',
 'birthdate': datetime.date(1947, 1, 10)}

And if we wanted to - again - create a pandas data frame from this same mock fake data:

df2 = []

for n in range(3):
    df2.append(list(faker.profile().values()))

df2 = pd.DataFrame(df2, columns=faker.profile().keys())
df2

There are many other options you can use to generate other fake data and also to tweak how some of the properties are generated. Again, the docs are very helpful indeed!

You May Also Like...