This project showcases a full data pipeline and dashboard built from a UK-based e-commerce dataset (2010–2011). It uses dbt for data modeling and transformation, DuckDB as the local warehouse, and Tableau Public for interactive visualizations.
- dbt-duckdb: SQL-based data modeling and transformation
- DuckDB: Local analytics database for storing model outputs
- Python: For reading DuckDB models and exporting to CSV
- Tableau Public: For designing and publishing dashboards
- Git + GitHub: Version control and project sharing
- Used a public e-commerce dataset with customer transactions from Kaggle
- Cleaned and stored CSV in a
/data/folder
- Set up a local dbt project using DuckDB adapter
- Created the following models:
stg_orders: Staging model to clean and cast raw CSV datafct_monthly_sales: Monthly revenue, quantity, and unique customersdim_customers: Total purchases, items bought, and revenue per customerfct_top_products: Top 10 products by total revenue and units soldfct_country_sales: Revenue and customer count by country
- Queried each dbt model using Python and
duckdb - Exported the final model outputs as CSVs to the
data/folder
- Connected to the exported CSVs in Tableau Public
- Built 3 key visualizations:
- Monthly Revenue Trend (Line Chart)
- Top Products by Units Sold (Bar Chart)
- Revenue by Country (Map)
- Combined all views into an interactive dashboard
- Published the dashboard to Tableau Public
- Revenue Trends: Sales spiked in November–December, likely due to holiday shopping.
- Top Products: Gift sets and vintage-themed items drove most of the revenue.
- Customer Behavior: A small portion of customers were highly active buyers.
- Geography: The UK led all countries in sales volume and revenue.
🔗 View the dashboard on Tableau Public
You can also find the full Tableau workbook:
ecommerce_dashboard.twbx