Design-an-AB-Test

View project on GitHub

Welcome to my “Design an A/B Test” Project Page!

I completed this project for Udacity’s Data Analyst Nanodegree “A/B Testing” course.

In this course, I learned how to do the following things:

  • Select metrics to evaluate a proposed change
  • Characterize and validate those metrics
  • Plan an appropriate duration for your experiment based on the number of samples needed and the expected risk
  • Sanity check the results to make sure everything went smoothly
  • Draw a conclusion based on the results
  • Recommend whether or not to launch the change

For this project, I considered an actual experiment that was run by Udacity. I fleshed out the experiment idea into a fully defined design, analyzed the results, recommended a decision, and proposed a high-level follow-up experiment. I also designed an A/B test, including a hypothesis, unit of diversion, and which metrics to measure.

A PDF file of the project can be found by clicking the “View on GitHub” link above.

Enjoy!