From Flat to Snowflake: Data modeling in Power BI
An exploration of different approaches to data modeling.
The Project
This project explores the evolution of analytical data modeling in Power BI through a progression of increasingly scalable schema architectures using the Adventure Works dataset.
Starting from flat schemas and moving toward Star and Snowflake models, the project focuses on transforming raw transactional data into structured, analysis-ready systems for business intelligence and reporting.
Along the way, the workflow covered data cleaning, duplicate removal, relational merging, hierarchy normalization, and relationship optimization to better understand how modeling decisions impact scalability, maintainability, and analytical performance.
Key points
Built Flat schemas from single and multiple data sources
Cleaned and validated transactional sales data
Designed a Star schema with fact and dimension tables
Normalized dimensions into a Snowflake schema
Configured relationships, cardinality, and filter direction
Explored scalability and performance-oriented modeling decisions
Prepared structured datasets for business reporting and BI workflows