Independent research: GraphRAG makes AI agents 80% more truthful | Read the report

NODES 26 — November 12, 2026

A Metamodel for Neo4j: How Deutsche Bahn Describes Meaning and Rules Within Neo4j

Session track: Data Intelligence

Session time:

Session description:

Neo4j’s schema-free property graph model is powerful, but it leaves important semantic questions open: What does a label, property, or relationship type mean? Which properties should labels or types have, and which label combinations should relationship types connect? This talk introduces a lightweight metamodel and ontology framework that defines these meanings directly inside the Neo4j database, keeping semantics available, synchronized, and compatible with evolving or non-conforming graphs. [1] The approach follows a practical “maybe, should, could” mindset rather than a rigid logic system, inspired by duck typing and open-world thinking. It also shows how our opensource GraphEditor supports the metamodel through semantic browsing, editing, visualization, parameterized queries, full-text search, persistent graph perspectives, and styling. Finally, the talk compares this approach with RDF 1.2. [1]: https://github.com/dbsystel/grapheditor/blob/main/docs/metamodel.md

Speaker

photo of Jörg Baach

Jörg Baach

Knowledge Architect, DB Systel

Jörg Baach, knowledge architect at DB Systel, has been working on graph-related projects for over a decade. Before that he was a web developer for more than 25 years.