This project analyzes an A/B test for a landing page to determine if the new version leads to a higher conversion rate compared to the old one. The test uses statistical analysis to validate business decisions.
- Name: ab_data.csv
- Size: ~294k rows
- Features: user ID, timestamp, group (control/treatment), landing page, conversion status
- Data cleaning and filtering mismatched group-page pairs
- Conversion rate analysis
- Hypothesis testing using Z-test for proportions
- Business recommendation based on statistical results
- Control group conversion rate: 12.04%
- Treatment group conversion rate: 11.88%
- P-value = 0.1897 → not statistically significant
- Conclusion: No measurable benefit from the new landing page
- Python
- Pandas
- statsmodels
- Google Colab / Jupyter Notebook
Prakash Sharma