API reference

This section provides the API reference for the Python modules that make up the data synthesis workflow. For more in-depth information, and synthpop-py in practice, see the User Guides and Examples.

The reference is organised by functionality to make it easier to find the components you need. The Synthesiser class provides the main interface for generating synthetic data, while the Data Preparation components support the encoding and handling of missing values required during synthesis. The available Methods implement different approaches to data synthesis, and the Utility Metrics can be used to assess the similarity and utility of synthetic data.

Additional documentation is provided for Plotting to help visualise results and internal Reproducibility to support consistent and repeatable synthesis workflows.

Use the sections below to explore the available classes, methods, and utilities, including their parameters, return values, and usage details.

Table of contents

Synthesiser class

• Synthesiser
  • Synthesiser
    • Synthesiser.fit()
    • Synthesiser.generate()

Data Preparation

• Encoding
  • PCAEncoder
    • PCAEncoder.fit()
    • PCAEncoder.transform()
  • MeanEncoder
    • MeanEncoder.fit()
    • MeanEncoder.transform()
• Missing Value Handling
  • MissingValuePredictor
    • MissingValuePredictor.prepare_data_for_fit()
    • MissingValuePredictor.post_synth_transform()
    • MissingValuePredictor.clone()
  • ReplaceMissingWithValue
    • ReplaceMissingWithValue.prepare_data_for_fit()
    • ReplaceMissingWithValue.post_synth_transform()
    • ReplaceMissingWithValue.clone()

Methods

• Base synthesis
  • BaseSynthMethod
    • BaseSynthMethod.fit()
    • BaseSynthMethod.transform()
    • BaseSynthMethod.get_feature_names_out()
• CART synthesis
  • CartMethod
    • CartMethod.fit()
    • CartMethod.transform()
    • CartMethod.get_feature_names_out()
    • CartMethod.fit_transform()
  • tune_cart()
  • TreeClassifierMethod
    • TreeClassifierMethod.transform()
    • TreeClassifierMethod.fit()
    • TreeClassifierMethod.fit_transform()
  • TreeRegressorMethod
• Copy synthesis
  • CopyMethod
    • CopyMethod.fit()
    • CopyMethod.transform()
    • CopyMethod.get_feature_names_out()
• Sample synthesis
  • SampleMethod
    • SampleMethod.fit()
    • SampleMethod.transform()
    • SampleMethod.get_feature_names_out()

Utility Metrics

• Pairwise S_pMSE
  • pairwise_spmse()

Plotting

• Univariate Distribution
  • plot_univariate_distributions()
• Pairwise S_pMSE
  • plot_spmse()

Reproducibility

• Reproducibility
  • RandomStateManager
    • RandomStateManager.set_root_seed()
    • RandomStateManager.create_instance_seed()
    • RandomStateManager.create_rng()