Designing and Implementing Semantic Data Layers, by Dave Wells
Data has a meaning problem. Semantic data layers solve it.
Ontology and Taxonomy
What Is Data Semantics?
The Roles of Semantics in Data Management
Why Semantics Is Needed Now
Five Types of Semantic Data Layers
Architectural Components of a Semantic Layer
Implementation Components of a Semantic Layer
Semantic Data Modeling Begins with the Domain
Knowledge Graphs
Property Graphs
The Semantic Data Modeling Process
Ontology: Concepts, Relationships, and Definitions
Properties in the Semantic Model
Taxonomy Analysis
Putting it All Together
The Semantic Model as Architectural Anchor
Semantic Layers in Architectural Context
The Enterprise Layer as Architectural Foundation
Domain Layers and Operational Data Management
Domain Layers and Data Products
Integration Layers: From ETL Coding to Semantic Mapping
The Enrichment Layer: Preserving Meaning Through Analytics Pipelines
Consumption Layers and Consumer Interfaces
From Implicit to Explicit Meaning
From Architecture to Software
Implementation Approaches
Four Implementation Patterns
Which Patterns Fit Where?
Choosing Patterns Based on Need
Semantic Layer Components in Operation
Semantic Layer Engineering
Four Architectural Patterns
Where to Start
How Much Coverage Is Needed?
Sequencing the Work
The Parallel Path: Semantic Modeling and Architectural Evolution
Designing for Evolution
Principles for Your Architectural Journey
Applying the Principles
Assessing the Current State
Best Practices for Data Semantics
The Broader Significance
Where to Begin
Semantics, applied to data, is the discipline of making meaning explicit, consistent, and durable. Semantic data layers are the architectural expression of that discipline. This book shows you how to build them.
Data management has always been messy. Inconsistent metrics, brittle integrations, data silos, and confusing dashboards persist even after heavy investment in data lakes, warehouses, pipelines, BI tools, metadata catalogs, and modern platforms. At the core, these are not technology problems; they’re problems of meaning. More infrastructure doesn’t fix a meaning problem. It expands it. AI workloads add new urgency: systems that act autonomously on data they don’t fully understand compound the cost of every inconsistency.
Semantic data layers turn scattered data definitions into shared, stable, reusable business meaning that is independent of the physical structures that hold the data. The result is strong data interoperability, reliable analytics, improved self-service access, reduced technical debt, and a practical foundation for trustworthy AI.
Building semantic layers well requires three things in the right order. First, architecture: understanding the five types of semantic layers (enterprise, domain, integration, enrichment, and consumption) and where each belongs in your data management world. Next comes design: ontology, taxonomy, semantic models, knowledge graphs, and property graphs. This is the conceptual work that captures how your business actually understands its data. Semantic models that business people can read and validate are more durable than technically precise models that only engineers can interpret. Then comes implementation: building semantic layers as software through APIs, data products, data contracts, data virtualization, and schema registries.
Architecture first, next design, and then implementation. This book covers all three. Written for data architects, data engineers, data leaders, data governance professionals, analytics teams, and the emerging Semantic Layer Engineer, it is a practical guide to reducing data friction, improving data quality, increasing reuse, and enabling self-service analytics across a complex enterprise.
When data has shared meaning, systems connect more easily, people trust the answers, and AI has the context it needs to reason responsibly.
Dave Wells is a data management consultant and educator with experience across a broad spectrum of data management processes and practices. As a consultant, he provides advice, direction, and guidance for data architecture, data quality, data governance, data integration, and data interoperability. As an educator, he is the Director of Education, an instructor at eLearningCurve, and an instructor for a variety of courses at Dataversity. Several decades of information systems, data management, and business management experience give Dave a well-balanced perspective about the synergies of business, information, data, and technology. Knowledge sharing and skills building are Dave’s passions, carried out through consulting, speaking, teaching, and writing.
Please complete all fields.
