Hackathon Showcase

LinkedInActivitySummary

Use the LinkedIn Portability API to see your activity. Create a GraphDB/RAG to search the posts you liked, shared etc as your personal knowledge base.

1 member

Project Description

For now the prototype works locally with individual python scripts.

Call the LinkedIn Portability API, grab my changelog. Extract some statistics, and build a json file with nodes (People, Posts, Comments) and relationships (has commented, reactions, …) [This part I had started before the hackathon as a side project, but I finalized at the beginning of the hackathon.]
Build a Neo4j graph using Cypher & AuraDB
Index the nodes with Content using neo4j-graphrag-python and Gemini models & embedders through VertexAI
Run a CLI to query the Graph RAG.

All this is mostly vibe-coded with Cursor Agents.

Team

Anne-Marie Tousch

Products & Tools

AI Tinkerers Google Neo4j

Additional Links

https://github.com/annemariet/linkedin-graph-rag

Code (WIP)

Summarizing URL...