For hundreds of years, alchemists tried to figure out how to turn ordinary metals into gold. The modern version of alchemists, let's call them Data Alchemists, try to turn raw, messy data into a high-quality, business-valuable asset.

A medallion architecture represents the most widely adopted pattern for organizing the data in modern data platforms, yet it's hard to believe that something so fairly simple by its nature may cause so much confusion and debate in the data industry

This session serves as a definitive guide to designing and implementing medallion architecture the right way! We'll start by explaining the core principles and recommended practices for implementing a medallion design pattern in a data platform solution.

We'll then examine different approaches to crafting medallion layers and answer the following questions: Should your bronze layer contain anything in addition to the raw data? Should you model your silver layer as a 3rd-normal-form-like-tables, or keep them as a "1 to 1" copy of the bronze layer? Star schema vs One-big-table in the gold layer? Should you perform calculations in the silver or in the gold layer? Or maybe in both?

Theory without practice is like a soup without salt - so, I'll take you on a ride to build the entire medallion architecture in Microsoft Fabric using a fictitious company that is just embarking on its data modernization journey. I'll teach you how to become their Chief Data Alchemist!