A/B Testing - Control vs Test Group Analysis
This project demonstrates a simple A/B testing process using Python. We check whether there is a statistically significant difference between the purchasing behaviors of two groups: Control Group and Test Group.
📄 Project Content • Data Reading (pandas) • Hypothesis Setting • Normality Test (Shapiro-Wilk Test) • Variance Homogeneity Test (Levene Test) • Independent Two-Sample T-Test (equal variance assumed) • Final Decision and Interpretation
🛠 Tools & Libraries • Python 3 • pandas • numpy • matplotlib • scipy.stats
📊 Dataset
The dataset contains: • Impression (Number of Ad Views) • Click (Number of Clicks) • Purchase (Number of Purchases) • Earning (Revenue)
Two groups were analyzed separately: • Control Group (Old Ad Strategy) • Test Group (New Ad Strategy)
🧪 Hypotheses • H0 (Null Hypothesis): The mean Purchase values of the Control and Test groups are equal. • H1 (Alternative Hypothesis): The mean Purchase values of the Control and Test groups are different.
📈 Conclusion • Both groups were normally distributed and had equal variances. • Based on the t-test results, the p-value > 0.05. • Therefore, we fail to reject the null hypothesis. • Result: There is no statistically significant difference between the purchase means of the control and test groups.