cluster-experiments: A Python library for end-to-end A/B testing workflows
About this talk
Schedule
Room:203
Date: April 24
Time: 12:00–12:45
Abstract
In this talk, we introduce cluster-experiments, a Python library designed to facilitate end-to-end A/B testing workflows, including power analysis, experiment analysis, and variance reduction techniques.
Prerequisites
python, basic stats knowledge
Description
We will go through the main techniques for mde analysis (simulation based and using Central limit theorem), how to do variance reduction in mde analysis and how to analyse experiments with the same library.
Speakers
David Masip
David Masip is a Data Science Manager at Glovo. He works in AB testing, causal inference and ML, and maintains two python libraries (cluster-experiments: a library for end-2-end analysis of AB tests, including experiment design; pytest-slow-last: a pytest plugin that makes slower tests run in the end).