This repository features A/B test analysis for a fast-food restaurant's menu launch. It covers data preparation, exploratory analysis, statistical tests, modeling, and visualizations. Promotion A, with high sales and consistent growth, is recommended for a nationwide launch. Ongoing monitoring is advised. Store Sales and A/B Test Analysis

The repository includes a Jupyter notebook walking through the data preparation, exploratory analysis, statistical testing, predictive modeling, and evaluation of results. Interactive visualizations are generated using Plotly and Bokeh.

The analysis compares performance metrics across three promotional strategies tested in a sample of restaurants over a four week period. Key techniques include linear regression, and time series forecasting.

Based on having the highest total sales and most consistent weekly growth, Promotion A is identified as the best performing strategy. The recommendation is to launch the new menu item nationwide using Promotion A.

Ongoing monitoring is advised to ensure continued positive performance once implemented. Adjustments to the promotional mix may be required based on the data insights.

This repository provides a template for a comprehensive A/B testing analysis covering the full process from data to visualization to modeling and recommendations. The documented methodology and interpreted results offer guidance for data-driven decision making.

The repository is organized as follows:

• Notebooks: Contains Jupyter notebooks for data analysis and modeling.

  • sales_analysis.ipynb: Analyzes store sales data, including market sizes, promotions, and weekly sales trends, Conducts A/B test analysis,
  • compares promotional strategies, and provides recommendations.

• Data: Stores input data and datasets used in the analysis.

  • dataset.csv:Input data with weekly sales figures for the new menu item under each promotional strategy.

• Outputs: Contains generated plots, visualizations, and result files.

• Store Sales Analysis:

  • Explores weekly sales trends, market sizes, and promotions.
  • Visualizes data using Seaborn, Plotly, and Bokeh.
  • Conducts time series forecasting for sales prediction.

• A/B Test Analysis:

  • Compares three promotional strategies (A, B, C) over a four-week period.
  • Recommends Promotion A for a nationwide launch based on performance metrics.

• Promotion A had the highest total sales.
• Weekly sales under Promotion A showed steady growth.
• Statistical tests confirmed Promotion A's effectiveness.
• Linear regression predicted a 15% sales lift for Promotion A.
• Forecasting predicts continued positive performance under Promotion A.

• Launch the new menu item nationally using Promotion A.
• Continuously monitor key performance indicators (KPIs).
• Be prepared to adjust promotional strategies based on data insights.

• Finalize a detailed launch plan for the national roll-out.
• Develop a real-time performance tracking dashboard.
• Optimize the marketing mix based on ongoing results.

• Mohammad Ghaffaripour