Modern Graph Databases
summary
Graph databases remain a niche but are growing for analytics and AI: Neo4j runs replicated graphs with trillions of edges and billions of nodes and avoids partitioning because graph partitioning is NP‑hard, while embedded analytical engines like KuzuDB load CSV/Parquet/JSON into memory and run Cypher for fast pattern queries. PuppyGraph translates Cypher/Gremlin over relational stores on the fly. Demos show cycle detection for money‑laundering (5‑hop matches) and building an ontology from RottenTomatoes with Gemini Flash for Graph RAG; live demos suffer Docker hiccups but ultimately load and query the graph.
AI-written summary · about
chapters (13)
- 0:00🚀 Build Stuff Conference – Get Your Tickets!
- 0:30Introduction: Why Graph Databases Are Different Today
- 3:10When Data Becomes “Graph-Worthy”
- 6:20Property Graphs vs RDF (Semantic Graphs)
- 10:40Querying Graphs: Cypher & Pattern Matching
- 15:20Why Graph Databases Are Still a Niche
- 18:30Neo4j & Challenges of Scaling Graph Systems
- 22:40Analytical Graph Databases: KuzuDB & In-Memory Processing
- 27:10Real Use Case: Fraud Detection with Graph Cycles
- 31:40Query Engines (PuppyGraph) & Graph on Top of SQL
- 36:10Graph + AI: Introduction to Graph RAG
- 39:20Building Ontologies with LLMs & Entity Extraction
- 41:30Final Thoughts: The Future of Graph Data & AI
related talks
26mGraph Databases will change your (freakin') life
01h 00mWho’s Afraid of Graphs?
29mSo... Graph Databases
32mUnderstanding Complexly Connected Data using Graph Databases
56mWhat are Graph Databases and Why should I care?
55mA Practical Guide to Graph Databases
41mGraphGen: Conducting Graph Analytics over Relational Databases
46mGraph Databases: Talking about your Data Relationships with Python