Automated A/B Testing Pipeline

Experimentation Python & SQL Statistical Analysis

An end-to-end automated pipeline for experimentation, handling data ingestion, significance testing, and result visualization.

The Challenge

Experimentation data collection and analysis were largely manual processes. Analysts spent significant time cleaning data and calculating p-values manually, which delayed test outcomes and created a dependency bottleneck.

Impact: This manual workload prevented the analytics team from focusing on high-value experiment design and strategy, limiting the volume of tests the business could run.

The Solution

I built an automated A/B testing framework that streamlined the entire lifecycle of an experiment.

Pipeline Automation: Built an end-to-end pipeline handling everything from data ingestion to result reporting.
Statistical Integration: Integrated statistical significance testing (t-tests, chi-square) directly into Tableau dashboards. This allowed stakeholders to see "Statistically Significant" flags instantly without manual calculation.
Self-Service: Enabled product managers to view experiment results in near real-time via the dashboard, removing the need for ad-hoc analyst requests.

The Impact

Throughput: Improved test reliability and throughput, allowing the team to run more concurrent experiments.
Efficiency: Freed up analyst bandwidth, shifting focus from data preparation to experiment design.
Speed: Significantly reduced the turnaround time for test results.