LLM-Assisted Data Exploration With the Neo4j Runway Python Library
summary
A Python library that uses Runway v0.14 beta plus GPT‑4o‑mini and GPT‑4o to turn CSVs into a Neo4j graph: supply a YAML data dictionary, run three async discovery calls that output unique IDs, candidate labels and relationships, then auto-generate a validated graph model (pydantic + instructor) and ingestion config. Ingests into Neo4j and visualizes regions, subregions, countries and shared currency nodes; includes a graph EDA module that reports node counts, disconnected nodes and high‑degree nodes. Targets an 80% functional graph, not production ready.
AI-written summary · about
chapters (8)
- 0:00Introduction to Runway
- 1:27Package architecture
- 2:56Loading source data
- 5:32Data discovery
- 7:02Modeling the graph
- 10:12Ingesting data
- 11:25Graph exploratory analysis
- 13:39Summary and conclusions
related talks
17mWorking with Neo4j with Django neomodel library
22mA Pythonic Tour of Neo4j and the Cypher Query Language
49mWarming up to Graphs
28mNeo4j for Beginners: Pros and Cons, Based on Social Networks
34mAre they ready? A quick review of graph databases and Python
46mGraph Databases: Talking about your Data Relationships with Python
36mGraphRAG Explained 🔥 Building Smarter LLMs
32mNever use a graph database!