Comprehensive Guide to Big Data Modeling and Data Warehouse Design

Chapter 1 introduces the need for data modeling in big‑data environments, emphasizing structured classification, cost reduction, efficiency improvement, and data quality, and contrasts OLTP (transaction‑oriented, 3NF) with OLAP (analysis‑oriented) systems.

It then surveys typical warehouse modeling approaches such as ER modeling, dimensional modeling (star and snowflake schemas), Data Vault, and Anchor models, highlighting their purposes and trade‑offs.

Chapter 2 describes Alibaba's data integration and management framework, detailing the layered architecture of ODS (operational data store), CDM (common dimension model) with DWD (detail) and DWS (summary) layers, and ADS (application data), together with principles of high cohesion, low coupling, cost‑performance balance, and naming consistency.

Chapter 3 focuses on dimension design, explaining basic concepts (facts vs dimensions, attributes, primary keys), design steps (selecting dimensions, defining granularity, identifying attributes), consistency and integration strategies, hierarchical and recursive dimensions, behavioral and multi‑value dimensions, and special cases such as micro‑dimensions.

Chapter 4 covers fact‑table fundamentals, classifying transaction, periodic snapshot, and cumulative snapshot fact tables, outlining design principles (granularity declaration, completeness, additivity, null handling, degenerated dimensions), and comparing single‑transaction versus multi‑transaction fact tables, including aggregation strategies, storage considerations, and implementation patterns used at Alibaba.