Analyze A/B Test Results
This project was completed through Udacity's Data Analyst Nanodegree certification.
Overview
The project conducted A/B testing of user conversions on an old and new webpage.
Procedure:
- Read data into a pandas Dataframe
- Handle Mismatched and duplicate records.
- Preform hypothesis, bootstrapping, Z-test and logistic regression.
- Determine if our P values are significant enough for further investigation.
Tech
- Python, Numpy, Pandas, Matplotlib, StatsModels
- LaTex
- Jupyter Notebook
Key Findings
- None of the variables have significant p-values.
- We will fail to reject the null.
- Not enough evidence to claim interaction between country and page received that on the basis of country there are more conversions.
- Everything appears to be working just fine, don’t fix what is not broken.