International E-Commerce Sales Analysis

Project Objective

This project simulates the role of a Data Analyst in an international e-commerce company. The goal is to analyze sales, customers, products, and countries to answer key business questions such as:

• Who are the top customers?
• Which products are the best sellers?
• Which countries generate the most revenue?
• What are the sales trends over time?

Dataset

Source: UCI Machine Learning Repository – Online Retail Dataset (UK)

• Real transactions from a UK-based online retailer (2010–2011)
• Fields: InvoiceNo, StockCode, Description, Quantity, InvoiceDate, UnitPrice, CustomerID, Country
• Format: Provided as .xlsx, converted to PostgreSQL and pandas DataFrame for analysis

Project Architecture & Tools

The project follows a complete data analysis pipeline, using industry-standard tools:

1. Data Import: Loaded the dataset into a PostgreSQL database (SQL)
2. Data Exploration & Cleaning: Wrote SQL queries for initial exploration and cleaning
3. Data Extraction: Imported query results into Python using pandas and SQLAlchemy
4. Data Cleaning & Transformation: Performed further cleaning, transformation, and aggregation in pandas
5. Visualization: Created insightful visualizations using matplotlib
6. Export: Exported results to CSV for reporting and documentation
7. Documentation: Documented the project with a README and visualization images

Business Questions Answered

The analysis addresses real-world business questions, including:

Sales

• How many orders were placed each month?
• What is the total and average revenue per country?
• What are the top 10 most sold products?

Customers

• Who are the top 10 customers by total spending?
• Which customers purchase most frequently?
• Which customers have the highest average order value?

Products

• Which products are most frequently returned (negative quantity)?
• Is there seasonality in product sales?

Example Visualizations

The project includes three visualizations:

• Top 10 countries by total revenue
• Top 10 most sold products
• Distribution of sales among top products

Technologies Used

• SQL (PostgreSQL) for data storage and querying
• Python (pandas, sqlalchemy, matplotlib)
• CSV/Excel for data export

Repository Structure

• dataCleaning.py: Data cleaning and quality checks
• dataVisualization.py: Data extraction, analysis, and visualization
• db_config.py: Database connection configuration
• analysis_queries.sql: SQL queries for exploration and aggregation
• OnlineRetailCSV.csv / OnlineRetail_cleaned.csv: Raw and cleaned datasets
• Visualization images:
  • Top 10 Most Sold Products.png
  • Distribution of Top 10 Most Sold Products.png
  • Total Revenue - Top 10 Countries.png
• README.md: Project documentation