What is an Ontology? Ontology Vs. Data Model Explained for Beginners
An ontology is a written definition of the things your business talks about and how they relate, in a form software can reason over. Here's how it differs from a taxonomy, a data model, a knowledge graph and a semantic layer. One shoe retailer carries the whole episode. We climb from a glossary to a taxonomy to an ontology, then map one customer who exists in three systems with three different keys onto a single class. You'll see how ontologies are written down (RDF, RDFS, OWL and SKOS), how a reasoner infers a fact nobody stored, and where the line falls between an ontology and a knowledge graph or semantic layer. Then we look at why data platforms are shipping ontologies for AI agents (Palantir, Microsoft Fabric IQ, Databricks Genie and Snowflake semantic views), and finish with a simple rule for deciding whether you need one. ⏱️ Chapters 0:00 - What is an ontology? 0:51 - Glossary, taxonomy, ontology: the ladder 2:02 - Ontology Vs. data model 3:46 - Writing it down: RDF, OWL and SKOS 5:05 - Inference: facts nobody stored 5:29 - Ontology Vs. knowledge graph 6:42 - Ontology Vs. semantic layer 8:01 - Why platforms ship ontologies for AI agents 9:16 - Action types: checking what an agent may do 9:39 - Do you need an ontology? 11:03 - Recap: the ladder side by side 🔗 Links W3C OWL: https://www.w3.org/OWL/ W3C RDF: https://www.w3.org/RDF/ W3C SKOS: https://www.w3.org/2004/02/skos/ Palantir Ontology: https://www.palantir.com/docs/foundry/ontology/overview/ #ontology #whatisanontology #ontologyvsdatamodel #knowledgegraph #semanticlayer #taxonomy #datamodeling #RDF #OWL #Palantir #MicrosoftFabric #Databricks #AIagents #dataengineering