Modelling for Data Engineers
Data modelling is the design of tables so analytics queries are fast and understandable. For data engineers it means dimensional modelling with facts and dimensions, star schemas, slowly changing dimensions, and the modern layered approach of staging, intermediate and mart models used in the warehouse.
Stage: Data Modelling and Warehousing · Part of the Data Engineer roadmap · Last reviewed 19 November 2026
What to learn in Modelling
- Design a star schema with a fact table and conformed dimensions
- Choose the grain of a fact table before adding any columns
- Implement a type 2 slowly changing dimension with valid_from and valid_to dates
- Organise models into staging, intermediate and mart layers
- Decide when a wide denormalised table is better than normalised tables
Modelling interview questions
What is the grain of a fact table and why does it matter?
The grain states what one row represents, such as one order line or one daily account balance. Fixing it first prevents double counting, because every measure and dimension key added later must be true at exactly that level of detail.
What is a slowly changing dimension type 2?
Type 2 keeps history by inserting a new dimension row when an attribute changes, with a surrogate key and effective date range or current flag. Facts then join to the version that was valid at the time, so past reports do not change.
Where Modelling fits in the Data Engineer roadmap
Modelling belongs to the Data Modelling and Warehousing stage (intermediate level), which takes about 4 hours. Make data usable.
Other units in this stage
Back to the full Data Engineer roadmap