Welcome to ABacus documentation!

ABacus is a Python library developed for A/B experimentation and testing. It includes versatile instruments for different experimentation tasks like experiment design, sample size determination, results evaluation, visualisation and reporting.

Important features

• Experiment design: type I/II errors, effect size, sample size simulations.

• Groups splitting.

• A/A test and evaluation of splitter accuracy.

• Evaluation of experiment results with various statistical tests and approaches.

• Sensitivity increasing techniques like CUPED, CUPAC.

Note

This project is under active development.

User Guide

• Installation
• Experiment Initialization
• Experiment Evaluation
• Splitter
• MDE Researcher

Technical Documentation

• Auto A/B
• Splitter
• Resplitter
• MDE Researcher

Examples

For more details, see the examples.

Communication

Developers and authors:

• Vadim Glukhov

• Egor Shishkovets

• Dmitry Zabavin