Create fake data with Faker in Python

Eric Callanan — Sunday, 28 Apr 2024

Fake data is useful in a number of scenarios such as in testing applications or libraries or in demonstrating functionality without any privacy or security concerns. While it is possible to create fake data manually, this can be time consuming and there is a better way. Faker is a library that makes it easy to generate fake data, such as names, phone numbers, emails, and other types of data.

There are many uses of fake data as listed below. In this article, I'll detail the basic use of Faker to create fake data, as well its use to populate a Pandas DataFrame and generate a chart.

Unit testing
Populating databases
Creating sample Reports
Simulating financial transactions
Load testing for APIs
Prototyping and design
Testing search functionality
Educational environments
Anonymizing data

Faker installation and use

Faker can be installed with pip install

1pip install faker

Simply import faker and initialize a faker generator, which can be used to create fake data by calling methods on the fake instance. The following generates a fake name by calling the name method on the instance of faker.

1from faker import Faker
2
3fake = Faker()
4
5# fake name
6fake.name()
7
8# 'Stefanie Brown'

Generate fake name using Faker

Show all the Faker methods

The following code lists all the methods on the instance of faker. This shows that there are 297 methods for creating all kinds of data.

 1methods = [x for x in dir(fake) if not x.startswith('_')]
 2print(f"number of methods = {len(methods)}")
 3# number of methods = 297
 4
 5
 6# print(methods)
 7for index in range(0, len(methods), 4):
 8    print(methods[index].ljust(35, " "), end="")
 9    try:
10        print(methods[index+1].ljust(35, " "), end="")
11        print(methods[index+2].ljust(35, " "), end="")
12        print(methods[index+3].ljust(35, " "))
13    except IndexError:
14        break

Methods on Faker:

Methods	methods (contd.)	methods (contd.)	methods (contd.)
aba	add_provider	address	administrative_unit
am_pm	android_platform_token	ascii_company_email	ascii_email
ascii_free_email	ascii_safe_email	bank_country	basic_phone_number
bban	binary	boolean	bothify
bs	building_number	cache_pattern	catch_phrase
century	chrome	city	city_prefix
city_suffix	color	color_hsl	color_hsv
color_name	color_rgb	color_rgb_float	company
company_email	company_suffix	coordinate	country
country_calling_code	country_code	credit_card_expire	credit_card_full
credit_card_number	credit_card_provider	credit_card_security_code	cryptocurrency
cryptocurrency_code	cryptocurrency_name	csv	currency
currency_code	currency_name	currency_symbol	current_country
current_country_code	date	date_between	date_between_dates
date_object	date_of_birth	date_this_century	date_this_decade
date_this_month	date_this_year	date_time	date_time_ad
date_time_between	date_time_between_dates	date_time_this_century	date_time_this_decade
date_time_this_month	date_time_this_year	day_of_month	day_of_week
del_arguments	dga	domain_name	domain_word
dsv	ean	ean13	ean8
ein	email	emoji	enum
factories	file_extension	file_name	file_path
firefox	first_name	first_name_female	first_name_male
first_name_nonbinary	fixed_width	format	free_email
free_email_domain	future_date	future_datetime	generator_attrs
get_arguments	get_formatter	get_providers	hex_color
hexify	hostname	http_method	iana_id
iban	image	image_url	internet_explorer
invalid_ssn	ios_platform_token	ipv4	ipv4_network_class
ipv4_private	ipv4_public	ipv6	isbn10
isbn13	iso8601	items	itin
job	json	json_bytes	language_code
language_name	last_name	last_name_female	last_name_male
last_name_nonbinary	latitude	latlng	lexify
license_plate	linux_platform_token	linux_processor	local_latlng
locale	locales	localized_ean	localized_ean13
localized_ean8	location_on_land	longitude	mac_address
mac_platform_token	mac_processor	md5	military_apo
military_dpo	military_ship	military_state	mime_type
month	month_name	msisdn	name
name_female	name_male	name_nonbinary	nic_handle
nic_handles	null_boolean	numerify	opera
optional	paragraph	paragraphs	parse
passport_dates	passport_dob	passport_full	passport_gender
passport_number	passport_owner	password	past_date
past_datetime	phone_number	port_number	postalcode
postalcode_in_state	postalcode_plus4	postcode	postcode_in_state
prefix	prefix_female	prefix_male	prefix_nonbinary
pricetag	profile	provider	providers
psv	pybool	pydecimal	pydict
pyfloat	pyint	pyiterable	pylist
pyobject	pyset	pystr	pystr_format
pystruct	pytimezone	pytuple	random
random_choices	random_digit	random_digit_above_two	random_digit_not_null
random_digit_not_null_or_empty	random_digit_or_empty	random_element	random_elements
random_int	random_letter	random_letters	random_lowercase_letter
random_number	random_sample	random_uppercase_letter	randomize_nb_elements
rgb_color	rgb_css_color	ripe_id	safari
safe_color_name	safe_domain_name	safe_email	safe_hex_color
sbn9	secondary_address	seed	seed_instance
seed_locale	sentence	sentences	set_arguments
set_formatter	sha1	sha256	simple_profile
slug	ssn	state	state_abbr
street_address	street_name	street_suffix	suffix
suffix_female	suffix_male	suffix_nonbinary	swift
swift11	swift8	tar	text
texts	time	time_delta	time_object
time_series	timezone	tld	tsv
unique	unix_device	unix_partition	unix_time
upc_a	upc_e	uri	uri_extension
uri_page	uri_path	url	user_agent
user_name	uuid4	vin	weights
windows_platform_token	word	words	xml
year	zip	zipcode	zipcode_in_state
zipcode_plus4

Fake methods for Name

It can be seen from the list of all the methods on Faker that there are a number of methods for generating names. We can use dir to list all the methods for generating fake names. This lists 24 methods that have name in the method name. Not all of these are names for people

1[x for x in dir(fake) if not x.startswith('_') and x.find('name') > -1]

Methods in Faker with name int the method

Fake Names	Fake Names (contd.)	Fake Names (contd.)
color_name	cryptocurrency_name	currency_name
domain_name	file_name	first_name
first_name_female	first_name_male	first_name_nonbinary
hostname	language_name	last_name
last_name_female	last_name_male	last_name_nonbinary
month_name	name	name_female
name_male	name_nonbinary	safe_color_name
safe_domain_name	street_name	user_name

Methods for people names in Faker

Fake Names	Fake Names (contd.)	Fake Names (contd.)
first_name_female	name	last_name
last_name_female	name_nonbinary	last_name_nonbinary
name_male	name_female	user_name
first_name_male	first_name
last_name_male	first_name_nonbinary

Generate sample fake names

It can be helpful to see a sample of data from these methods to help identify the best method for the desired fake data. This code calls the methods with name in them to generate a sample for each of the methods. This is very useful to compare the output from the different names.

1for name_method in [x for x in dir(fake) if not x.startswith('_') and x.find('name') > -1]:
2    print(f"{name_method.ljust(23, ' ')} :   { eval(f'fake.{name_method}()') }")

Fake data generated for methods with name

Method Name	Fake value generated
color_name	Navy
cryptocurrency_name	Zcash
currency_name	Surinamese dollar
domain_name	miller-barber.com
file_name	ok.webm
first_name	Aaron
first_name_female	Patricia
first_name_male	Jackson
first_name_nonbinary	Ricky
hostname	web-01.hampton-jenkins.com
language_name	Bihari languages
last_name	Ray
last_name_female	Barker
last_name_male	Vega
last_name_nonbinary	Morales
month_name	August
name	Charles Hoover
name_female	Kathryn Levy
name_male	Timothy Bruce
name_nonbinary	Samantha Carr
safe_color_name	gray
safe_domain_name	example.com
street_name	Kevin Harbor
user_name	haneydaniel

Generate sample fake names using Faker

Regenerating the same fake data - Seed

It is sometimes necessary to generate the same fake data rather than generating random data every time. seed produces the same result when the same methods with the same version of faker are called. This can be required if fake data is being used in unit tests. This is easily achieved by setting a seed similar to random number generation.

 1Faker.seed(42)
 2for i in range(5):
 3    print(fake.name())
 4# Allison Hill
 5# Noah Rhodes
 6# Angie Henderson
 7# Daniel Wagner
 8# Cristian Santos
 9
10
11Faker.seed(42)
12for i in range(5):
13    print(fake.name())
14# Allison Hill
15# Noah Rhodes
16# Angie Henderson
17# Daniel Wagner
18# Cristian Santos

Localized fake data

A locale value can be passed into the constructor for Faker to create fake data in that locale.

1italian_fake = Faker('it_IT')
2for name_method in [x for x in dir(italian_fake) if not x.startswith('_') and x.find('name') == 0]:
3     print(f"{name_method.ljust(23, ' ')} =   { eval(f'italian_fake.{name_method}()') }")
4
5# name                    =   Lando Opizzi
6# name_female             =   Elvira Bonanno-Beccheria
7# name_male               =   Dott. Jacopo Bottaro
8# name_nonbinary          =   Ciro Parmitano

1japanese_fake = Faker('ja_JP')
2for name_method in [x for x in dir(japanese_fake) if not x.startswith('_') and x.find('name') == 0]:
3    print(f"{name_method.ljust(30, ' ')} =   { eval(f'japanese_fake.{name_method}()') }")
4
5# name                           =   山田 春香
6# name_female                    =   小林 知実
7# name_male                      =   橋本 稔
8# name_nonbinary                 =   斎藤 健一

1chinese_fake = Faker('zh_CN')
2for name_method in [x for x in dir(chinese_fake) if not x.startswith('_') and x.find('name') == 0]:
3    print(f"{name_method.ljust(30, ' ')} =   { eval(f'chinese_fake.{name_method}()') }")
4
5# name                           =   张春梅
6# name_female                    =   权建平
7# name_male                      =   贺华
8# name_nonbinary                 =   阎玉梅

Create localized Faker instance to generate localized fake data

Create a chart with fake data

Any of the methods on the instance of Faker can be used in a list comprehension to generate lists of fake values. These lists can be combined into a list of dictionaries and these dictionaries used to load data into a Pandas DataFrame. This can be a great way to quickly generate test data without having to risk exposing customer sensitive data. Once the data is loaded into a DataFrame it can be manipulated or charts created to develop data pipelines with the data.

1Faker.seed(42)
2DATA_SIZE = 500
3names = [fake.first_name() for i in range(DATA_SIZE)]
4ages = [fake.random_int(min=21, max=80) for x in range(DATA_SIZE)]
5states = [fake.state() for x in range(DATA_SIZE)]
6jobs = [fake.job() for x in range(DATA_SIZE)]
7
8data = {'name': names, 'age': ages, 'state': states, 'job': jobs}
9df = pd.DataFrame.from_dict(data)

Top 10 records from a DataFrame containing fake data

	Name	Age	State	Job
0	Danielle	60	Delaware	Surveyor, planning and development
1	Angel	55	Alaska	Surveyor, planning and development
2	Joshua	22	South Dakota	Landscape architect
3	Jeffrey	79	Rhode Island	Community education officer
4	Jill	46	Kansas	Paediatric nurse
5	Erica	58	Virginia	Manufacturing systems engineer
6	Patricia	57	Hawaii	Pharmacologist
7	Christopher	63	Maryland	Arts development officer
8	Robert	22	Oklahoma	Holiday representative
9	Anthony	26	Alabama	Tourism officer

 1mean_age_df = df[['state', 'age']].groupby('state').mean()
 2mean_age_df.rename_axis(None, axis = 0, inplace = True)
 3top_df = mean_age_df.sort_values(by=['age'], ascending = False).head(10)
 4
 5fig, ax = plt.subplots(figsize = (10,7), facecolor=plt.cm.Blues(.2))
 6ax.set_facecolor(plt.cm.Blues(.2))
 7
 8fig.suptitle('Top 10 States with highest average age',
 9             fontsize = 18,  
10             fontweight = 'bold')
11
12states = list(top_df.index)
13y_pos = np.arange(len(states))
14ages = list(top_df['age'])
15
16ax.barh(y_pos, ages, align='center')
17ax.set_yticks(y_pos)
18ax.invert_yaxis()
19ax.set_xlabel('Average Age', fontsize = 14)
20
21# Display highest on top
22ax.set_yticklabels(states, fontsize = 14)
23
24# Hide the right and top spines
25ax.spines['right'].set_visible(False)
26ax.spines['top'].set_visible(False)
27
28plt.show

Create bar chart from fake data loaded into a Pandas DataFrame

Conclusion

Fake data can easily be created with Faker in Python. I can't help asking - do we really need more fake data? I think we do, as it can be useful in a number of situations, such as unit testing or developing data pipelines without the need for real data. There is wide list of data types that can be generated with Faker as well as providing a custom data provider.

csv on Software Development Notes