Mixed Model Arts: The Building Blocks of Cross-Disciplinary Data Modeling

In 1993, the Ultimate Fighting Championship settled an old argument. Every martial art claimed to have the answer. Then practitioners from different disciplines competed directly. No single style had all the answers. The fighters who came to dominate learned across disciplines and used what worked. Data modeling is having its UFC moment. For decades, data modeling has evolved into five broad camps: relational, analytics, applications, ML/AI, and knowledge. Each developed powerful techniques for the problems it was trying to solve. Each also developed its own vocabulary, assumptions, and blind spots. Meanwhile, real-world data stopped respecting those boundaries and moved on. Your product catalog lives in JSON. Your analytics dashboard disagrees with your recommendation engine. Your knowledge graph describes concepts that your operational database represents differently. AI agents now read across documents, metrics, metadata, embeddings, and operational systems. No single camp has all the answers. Enter Mixed Model Arts. Written by Joe Reis, co-author of Fundamentals of Data Engineering, Mixed Model Arts takes a pragmatic approach to modern data modeling: learn from every tradition, understand the fundamental concepts beneath them, and apply the techniques that fit the problem in front of you. Featuring a foreword by Bill Inmon, the father of data warehousing. Inside, you’ll learn how to: Work across the five camps of data modeling: relational, analytics, applications, ML/AI, and knowledge Model the five forms of data: structured, semi-structured, metadata, unstructured, and ML/AI artifacts Use concepts that cut across every form: entities, identifiers, attributes, relationships, grain, time, and meaning Understand grain: one of the most important (and most frequently misunderstood) concepts in data modeling Model data for AI systems and agents: so machines can work with data in context rather than merely retrieve it Model the business itself: instead of simply reproducing source systems Navigate organizational reality: including the incentives, politics, and communication failures that derail modeling efforts Who this book is for Data and analytics engineers trying to understand inherited schemas nobody can explain Backend developers designing data structures that need to survive beyond the next feature Architects and staff engineers creating shared structure across teams and systems Data leaders dealing with systems that disagree about the same business ML and AI practitioners discovering that many AI problems are really data problems underneath The book is light on code and heavy on reasoning. Read it straight through, or follow the reading paths in the front matter based on what you need. Every chapter ends with a “Try This” exercise because modeling is something you practice, not something you memorize. Mixed Model Arts: Book One The real work happens in the gym