Retail Sales Data Analysis Project

Project Overview

This project performs Exploratory Data Analysis (EDA) on a retail transaction dataset to uncover customer behavior patterns and provide actionable business recommendations.

Dataset:
Retail Sales Dataset

Dataset Columns:

• transaction_id
• Date
• Customer_ID
• Gender
• Age
• Product_Category
• Quantity
• Price_per_Unit
• Total_Amount (Revenue)

Tools & Libraries Used

• Python
• Pandas : Data manipulation and analysis
• Matplotlib : Basic visualizations
• Seaborn : Statistical plots (heatmap, boxplots, etc.)
• SQL Server : For running the analysis
• Excel : Pivot tables, Dashboard
• Jupyter Notebook : For running the analysis

Actionable Recommendations

• Retention Program for big spenders (critical to protect revenue).
• Upselling:
  • Focus on Electronics category
  • Target Age 18–25 (highest upselling efficiency)
  • Run during peak months
• Cross-selling:
  • Focus on Clothing category
  • Target Age 26–50
  • Run during low-revenue months
• Slightly raise prices on premium items.
• Marketing:
  • Keep gender-neutral strategy.
  • Extra focus on Age 18–25.
• Maintain a balanced strategy, prioritizing upselling for higher impact.

Project Structure

Retail-Sales-Analysis/
│
├── Cleaning/
│   ├── Cleaning.py
│   └── Cleaning_Notebook.ipynb
│
├── EDA/
│   ├── EDA.ipynb
│   ├── SQL_files/
│   │   ├── create_table.sql
│   │   ├── load_data.sql
│   │   └── EDA(group by, segmentation).sql
│   ├── EDA_Customer_Queries.sql
│   └── EDA_Revenue_Queries.sql
│
├── visualization/
│   └── Retail Sales Dashboard.xlsx
│
├── data/
│   └── retail_sales_dataset.csv
│
├── docs/
│   ├── define_problem.md
│   ├── data_catalog.md
│   └── actionable_insights.md
│
├── README.md
└── LICENSE

Conclusion

By prioritizing upselling on high-value items (especially Electronics for younger customers) and protecting big spenders, while using cross-selling strategically on Clothing during slower periods, the business can achieve higher and more stable revenue growth.

---

Project completed on December 27, 2025