Demo Project: Ecommerce DBT-BigQuery Project

This project demonstrates how DBT can be used to model data to adhere to an Inmon DWH Architecture stored in Google BigQuery (BQ).

Tools used:

• Data Source: The Look Ecommerce BigQuery Public Data
  • Data Dictionary can be found on Kaggle
  • The data was loaded from a CSV to demonstrate a simple way of how to load CSV data into BQ.
• Exploratory Data Analysis Tool: Jupyter Lab, Pandas, DuckDB, SQL
• Data Loader: Python Script using BigQuery API
• Transform: DBT
• Visualiation [WIP]: Looker Studio

Problem Statement

The Look, a growing eCommerce clothing site, would like to make more data-driven decisions. They want you to prioritize helping the Sales & Marketing team and the Operations team.

After consultations with both teams, you identified the following.

Sales & Marketing

• Wants to get a better understanding of the company's revenue drivers on a month-to-month basis
  • What are our top selling products?
  • What's their biggest market?
  • Who are their customers?
  • Where do their customers come from?
• Is the company's profitability growing over time?

Operations

• Wants to know if they're keeping up with their orders.
• Which warehouses should they expand the capacity of?

Caveats & Observations

• The prescribed retail_price is always equal to the sale price
• The location of the user is the delivery location
• All Timestamps are in UTC
• Not all items in the inventory_items table can be found in the order_items table.