A/B Testing Dashboard

An interactive dashboard to understand statistical significance in A/B tests. Built with Streamlit.

What This Does

This tool helps you:

• Simulate A/B tests with adjustable parameters
• Understand statistical significance (p-values, confidence intervals)
• Learn when a result is real vs. random chance
• Make data-driven business decisions

Features

• Interactive Sliders - Adjust sample size and conversion rates in real-time
• Statistical Testing - Chi-Square test for significance
• Confidence Intervals - See the likely true conversion rate range
• Visualizations - Compare versions with charts and metrics
• Business Interpretation - Plain English explanations of results

Live Demo

Try the interactive dashboard: [Your Streamlit Cloud link - we'll add this in a moment]

How to Run Locally

1. Clone the repository:

git clone https://github.com/evasafi/ab-test-dashboard.git
cd ab-test-dashboard

2. Install dependencies:

pip install streamlit numpy pandas scipy matplotlib seaborn

3. Run the app:

streamlit run ab_test_dshboard.py

4. Open http://localhost:8501 in your browser

How It Works

The Scenario

• Version A: Current design (control)
• Version B: New design (treatment)
• Test if Version B truly converts better than Version A

The Statistics

• P-Value: Probability the result is due to random chance (< 0.05 = significant)
• Confidence Interval: Range where the true conversion rate likely falls
• Chi-Square Test: Statistical test to determine significance

The Interpretation

If p-value < 0.05: Version B is statistically significantly better If p-value ≥ 0.05: The difference could be random chance—need more data

Use Cases

• E-commerce checkout optimization
• Marketing campaign A/B testing
• Website design improvements
• Email subject line testing
• Any scenario where you need to compare two versions

Technologies Used

• Python - Data analysis and statistics
• Streamlit - Interactive web dashboard
• SciPy - Statistical testing
• Matplotlib/Seaborn - Data visualization
• NumPy/Pandas - Data manipulation

About

This project demonstrates skills in:

• Statistical analysis and hypothesis testing
• Data visualization
• Interactive web application development
• Business analytics and decision-making

Built for portfolio purposes as a Business & Data Analyst.