complexifier 0.3.2

A library to make data worse

Complexifier

Make your pandas even worse!

Problem

When teaching students to work with data, an important lesson is how to clean it.

The problem with this is that there are two types of datasets available on the internet:

1. Data that is good, but already cleaned
2. Data that is not cleaned, but is terrible and incomprehensible

Complexifier solves this problem by allowing you take the former and turn it into a better version of the latter!

Dependencies

• pandas
• typo
• random

Installation

complexifier can be installed using pip

pip install complexifier

Usage

Go to the documentation

Once installed you can use complexifier to add mistakes and outliers to your data

This library has several methods available:

create_spag_error

create_spag_error(word: str) -> str

Introduces a 10% chance of a random spelling error in a given word. This function is useful for simulating typos and spelling mistakes in text data.

introduce_spag_error

introduce_spag_error(df: pd.DataFrame, columns=None) -> pd.DataFrame

Applies the create_spag_error function to each string entry in specified columns of a DataFrame, introducing random spelling errors with a 10% probability.

add_or_subtract_outliers

add_or_subtract_outliers(df: pd.DataFrame, columns=None) -> pd.DataFrame

Randomly adds or subtracts values in specified numeric columns at random indices, simulating outliers between 1% and 10% of the rows.

add_standard_deviations

add_standard_deviations(df: pd.DataFrame, columns=None, min_std=1, max_std=5) -> pd.DataFrame

Adds between 1 to 5 standard deviations to random entries in specified numeric columns to simulate data anomalies.

duplicate_rows

duplicate_rows(df: pd.DataFrame, sample_size=None) -> pd.DataFrame

Introduces duplicate rows into a DataFrame. This function is useful for testing deduplication processes.

add_nulls

add_nulls(df: pd.DataFrame, columns=None, min_percent=1, max_percent=10) -> pd.DataFrame

Inserts null values into specified DataFrame columns. This simulates missing data conditions.

mess_it_up

mess_it_up(df: pd.DataFrame, columns=None, min_std=1, max_std=5, sample_size=None,min_percent=1, max_percent=10, introduce_spag=True, add_outliers=True, add_std=True, duplicate=True, add_null=True) -> pd.DataFrame

Adds all (or some) of the above methods. Really messes it up.