โฌ
๏ธ Back to home
โจ Project 3 โ E-commerce Data Analysis with Power BI (The Look Dataset)
๐งญ Context and Objectives
This project involved analyzing a comprehensive dataset simulating operations of an online fashion retail site ("The Look") to extract key insights on sales, customer behavior, product performance, and marketing effectiveness.
Objective: identify strategic levers using Power BI from multichannel data.
๐งพ Data Used
- Multi-table database:
- Orders, products, users, inventory, web events, marketing, etc.
- Source: BigQuery โ connected to Power BI Desktop
- Volume: tens of thousands of rows
- Visualization and modeling in Power BI
๐ Technical Skills Used
- BigQuery connection from Power BI
- Data cleaning via Power Query Editor
- Relational modeling in Power BI (model view)
- KPI calculations: revenue, return rate, AOV, conversion rate, marketing ROIโฆ
- Customer and product segmentation
- Interactive visualizations (maps, graphs, pivot tables)
๐งฎ Analytical Approach
- Understanding the business context: study of "The Look" e-commerce model
- Selecting key analysis axes: sales, customer behavior, product performance
- Data exploration:
- Sales analysis by category, period, and channel
- Customer segmentation by spending, location, behavior
- Product return rate and turnover
- Creating interactive dashboards
- Presenting recommendations to improve overall performance
๐ Key Insights
- Top 3 categories generate over 65% of annual revenue
- Product returns are concentrated on certain brands โ quality or sizing issues
- Loyal customers (3+ orders) have an average basket 40% higher
- Email campaigns yield twice the ROI of social media ads
๐ก Recommendations
- Optimize stock for high-turnover categories and reduce high-return products
- Strengthen targeted marketing for loyal customers
- Cut back on low-ROI channels (e.g., social media) in favor of email marketing
- Improve product descriptions for most returned items
๐ What I Learned
- Using Power BI in real-world scenarios with complex datasets
- Efficiently cleaning and modeling multi-source data
- Defining clear, business-oriented analytical objectives
- Presenting actionable and relevant recommendations quickly