Graph Database and Retrieval Augmented Generation
summary
Builds a RAG pipeline that actually stores PyData submissions in Neo4j, indexes a document node into a vector store, then runs both vector retrieval and Cypher-based queries via LangChain graph Transformer and OpenAI embeddings. Demo returns concrete counts — 65 submissions, 16 tutorials, 49 talks — and shows an agent that switches between Cypher QA and vector retriever using a custom retrieval query. Prototype runs end to end but hits token and rate‑limit pain points when retrieval returns too much context.
AI-written summary · about
chapters (23)
- 0:00Graph databases and RAG intro
- 3:21Traditional RAG limitations
- 5:28Why RAG is needed
- 6:53Graph databases enhance RAG
- 8:01Graph RAG demo setup
- 11:57Graph database basics
- 19:17Structuring knowledge graph data
- 23:26Loading data into graph
- 31:36Building graph RAG system
- 34:12Creating document embeddings
- 35:39Cypher QA chain queries
- 38:09Combining retrieval and Cypher
- 43:08Scraping conference data
- 46:00Generating knowledge graph
- 50:19Querying enhanced knowledge graph
- 52:10Evaluating RAG performance
- 55:33Updating knowledge graphs
- 58:46Retrieval queries add context
- 1:03:27Visualizing knowledge graph
- 1:09:19Graph RAG challenges
- 1:12:11Optimizing graph RAG cost
- 1:15:22Pruning knowledge graph
- 1:16:07Closing remarks
related talks
36mGraphRAG Explained 🔥 Building Smarter LLMs
39mEnhancing RAG with Neo4j Knowledge Graph
50mRetrieval Augmented Generation Beyond Toy Demos
01h 40mKnowledge Graphs & GraphRAG: Techniques for Building Effective GenAI Applications
15mAgentic GraphRAG: Simplifying Retrieval Across Structured & Unstructured Data
48mGoing Beyond Prompt Engineering
27mWhen Vectors Aren’t Enough: Unlocking AI’s Hidden Connections with Graphs
16mEnhancing Retrieval-Augmented Generation with GraphRAG Patterns in Neo4j