Building an Entity-Resolved Knowledge Graph: Why and How
summary
Shows a concrete pipeline that turns the ICIJ offshore/Panama Papers graph into an entity-resolved knowledge graph by integrating a sensing engine (Senzing) with a Neo4j-backed visualization (Linkurious/Linkprise), producing an extra “real‑world entity” node per cluster. Claims: 1.9M entities collapse to 10–73 clusters in examples, 1.884M singletons and ~30k duplicates (4.1% duplication rate). Demonstrates an entity-centric deduplication method that resolves false negatives from record‑pair matching, runs real-time at scale and remains explainable for human review.
AI-written summary · about
chapters (9)
- 0:00Introduction to entity resolution
- 0:48The challenge of duplicate data
- 2:42Benefits of resolution for risk
- 4:27Approaches to entity resolution
- 7:25The entity-centric architecture
- 9:32Benefits of the integrated solution
- 11:37Panama Papers case study
- 14:43Demonstration of insights
- 22:41Workflow automation and conclusion
related talks
01h 41mGraph Intelligence: Enhance Reasoning and Retrieval Using Graph Analytics
30mHow Graph Data Science Can Turbocharge Your Knowledge Graph
36mThe World's Largest Open Source Entity Graph
15mGraphs on the High Seas: UI/UX Considerations for Great Geospatial Graph Visualization
52mUsing Open Source Tech to Unearth the Stories of the Panama Papers
42mBuilding a knowledge graph for climate policy
28mCreating Enterprise Knowledge Graphs from Unstructured Data
25mBuilding the World’s Largest Identity Graph without a Graph DB