Simple pseudonymization of a column with name values in a pandas dataframe

simple_pseudonymization.py

#!/usr/bin/python3
# -*- coding: utf-8 -*-

# Requires pandas and numpy
#     pip3 install pandas numpy
# and the Faker package:
#     pip3 install Faker

from faker import Faker
import numpy as np
import pandas as pd

# example dataframe
df = pd.DataFrame({'name': {0: 'foo', 1: 'bar', 2: 'foo', 3: 'foobar', 4: 'foobar', 5: np.nan, 6: ''}})

print(df)
#      name
# 0     foo
# 1     bar
# 2     foo
# 3  foobar
# 4  foobar
# 5     NaN
# 6        

# pseudonymization of the name column using Faker
fake = Faker()

# empty string values as nan
df.replace('', np.nan, inplace=True)

# name replacement dictionary
name_replacements = {name: fake.name().lower().replace(" ", "") for name in df['name'].unique() if name is not np.nan}

# apply replacement
df.replace({"name": name_replacements}, inplace=True)

print(df)
#            name
# 0     traceymata
# 1   susansmithmd
# 2     traceymata
# 3  charlesrogers
# 4  charlesrogers
# 5            NaN
# 6            NaN