A/B Test: E Commerce

For this project, I will be working to understand the results of an A/B test run by an e-commerce website if they should implement the new page, keep the old page, or perhaps run the experiment longer to make their decision.

This project includes the following contents:

• Introduction
• Part I: Probability
• Part II: A/B test
• Part III: Regression
• Conclusion

Before Part I: Probability, I will perform data cleaning such as checking missing data, discrepancies between the columns etc.

Language and Package

This project is using Python3. The packages are used in this project including Numpy, Panda, random, matplotlib.pyplot, scipy.stats, and statsmodels.api.

Metadata of variables

1. ab_data.csv

Variable Name	Metadata
user_id	6-digit numbers
timestamp	string
group	string: control, treatment
landing_page	string: old_page, new_page
converted	numeric: 0:No, 1:Yes

2. countries.csv

Variable Name	Metadata
user_id	6-digit numbers
country	string: US, CA, UK

Methodology

• A/B Test
• Two-Proportion z Test
• Logistic Regression

Report

The reports are generated to three formats, .ipynb, .pdf and .html:

• Analyze_ab_test_results_notebook_jc.ipynb
• Analyze_ab_test_results_notebook_jc.pdf
• Analyze_ab_test_results_notebook_jc.html

Feedback from the reviewer

• Udacity Reviews