Graph database
Database optimized for graph data — nodes with relationships. Queries via path traversal instead of joins. Strong where relationships are first-class (social networks, fraud, knowledge graphs).
Two models: property graph (Neo4j, Memgraph, JanusGraph — nodes + edges with properties) and RDF (Apache Jena, AllegroGraph — semantic web, triples). Query languages: Cypher (Neo4j, now also ISO GQL 2024), Gremlin (Apache TinkerPop), SPARQL (RDF). Practical use cases: social network, recommendation, fraud detection, network topology, supply chain. Multi-model DBs (ArangoDB, OrientDB) do graph + document.