Online Retail Simulator

Generating synthetic product and sales data for testing causal inference pipelines

Building and validating causal inference pipelines requires realistic data—but using production datasets introduces privacy, compliance, and operational constraints.

Online Retail Simulator generates fully synthetic retail data for end-to-end testing of causal inference workflows. It simulates products, customers, sales, and conversion funnels while preserving key statistical and behavioral patterns found in real e-commerce systems.

Unlike generic data generators, the simulator supports controlled treatment effects, enabling teams to validate estimators, stress-test identifying assumptions, and compare causal models against known ground truth—before running causal analysis on production data.

Quick Start

pip install git+https://github.com/eisenhauerIO/tools-online-retail-simulator.git

from online_retail_simulator import simulate, load_job_results

job_info = simulate("config.yaml")
results = load_job_results(job_info)

products_df = results["products"]
metrics_df = results["metrics"]

Documentation

Guide	Description
Usage	Getting started with basic workflow
Configuration	All configuration options
Design	System design and architecture