Beyond the Silos Achieving Data Harmony With Graph Based ER
summary
End-to-end graph-based ER on Neo4j + GDS uses a single streaming Cypher query for candidate selection and ranking, combining node-attribute, graph-pattern and composite rules; scores multiply a normalized similarity (0–1) by a rule importance weight. Resolution runs WCC for broad clusters and a weighted hierarchical community algorithm for finer splits. Demo projects ~400 person nodes, writes component IDs as node attributes, and an incremental CDC on connected components updates only affected subgraphs (5-node rerun) in under a minute.
AI-written summary · about
chapters (8)
- 0:00Introduction to talk
- 0:50Why entity resolution matters
- 3:42Candidate selection and ranking
- 11:14Resolution with community detection
- 15:08Incremental resolution process
- 18:24Data modeling options
- 20:10Live demo walkthrough
- 28:23Key takeaways and conclusion
related talks
32mUnderstanding Complexly Connected Data using Graph Databases
49mModern Graph Databases
19mUnraveling the Web of Relationships: Understanding Graph Data
30mContext Engineering: Connecting the Dots with Graphs
01h 01mA Skeptics Guide to Graph Databases
24mGraph-Powered Job Matcher: From SQL Tables to Neo4j RAG in Production
15mAgentic GraphRAG: Simplifying Retrieval Across Structured & Unstructured Data
30mPerformant GraphRAG Improving Neo4j to Drive Autonomous AI ...