DataGenix

An advanced and robust library for generating synthetic datasets for machine learning and deep learning projects. Go from idea to prototype in seconds without data acquisition bottlenecks.

Installation

Install from PyPI (once published):

pip install datagenix

Or install directly from the repository:

git clone [https://github.com/yourusername/datagenix.git](https://github.com/yourusername/datagenix.git)
cd datagenix
pip install .

Ultimate Usage Example

Generate a complex, realistic dataset for a binary classification task with a single, intuitive command:

from datagenix import DataGenerator

generator = DataGenerator(seed=42)

df = generator.generate(
    num_rows=1000,
    numerical_whole=3,
    decimal=2,
    categorical=2,
    boolean=1,
    text=1,
    uuid=1,
    object_types=['name', 'email'],
    target_type='binary',
    missing_numerical=0.05,
    missing_categorical=0.1,
    correlation_strength=0.7,
    group_by='customer_id',
    num_groups=50,
    time_series=True,
    numerical_whole_range=(100, 999),
    add_outliers=True,
    outlier_fraction=0.02,
    text_style='review'
)

print(df.head())
print(df.info())

Advanced Features

• Target Generation: Automatically create a target column for binary, multi-class, or regression tasks that is logically correlated with the features.
• Missing Data: Inject missing values (NaN) into any feature type with precise fractional control (e.g., missing_numerical=0.1).
• Feature Correlation: Create linear dependencies between numerical features with adjustable correlation_strength.
• Grouped Data: Simulate real-world scenarios like customer data by grouping rows with a common ID using group_by and num_groups.
• Time Series: Generate a chronologically sorted timestamp column for time-dependent modeling.
• Outlier Injection: Introduce extreme values into numerical columns to test model robustness using add_outliers and outlier_fraction.
• Custom Ranges: Define exact (min, max) ranges for numerical columns.
• Text Styles: Generate varied text content like review, tweet, or standard sentence.