Home / Insights
Notes from the work.
Modeling connected data, uncovering relationships, and applying graph techniques to real business problems. Occasional Cypher.
Start here.
A book, an hour of conversation, and a worked example.
Graph Data Science with Python and Neo4j
Hands-on projects for visualization and analysis. Orange AVA, 192 pages.
Making invisible connections visible
Tim on graph data science consulting, the book, PRAC oversight work, and knowledge graphs for private equity.
Graph features for a classic classification problem
Using graph structure as features on the cervical cancer dataset, and what that buys you.
Articles
Written from engagements, not from a keyword list.
Entity Resolution Techniques: How to Collapse Duplicate Records Without Losing the Truth
Exact matching finds the easy duplicates and misses the expensive ones. Here is how deterministic rules, probabilistic scoring and graph structure fit together, and where each one stops working.
Read the article
GraphRAG in Practice: Giving Enterprise AI the Context Vector Search Misses
Vector search finds passages that look alike. It has no idea which ones are actually connected. That gap is where most enterprise AI pilots stall, and it is what GraphRAG exists to close.
Read the article
Graph Analytics for Fraud Detection: Find the Ring, Not the Transaction
A rules engine asks whether this transaction looks wrong. Organised fraud is built so that no single transaction does. The signal is in what the accounts share, and that is a graph question.
Read the articleSee how graph intelligence actually works.
Practical notes on modeling connected data, uncovering relationships, and applying graph techniques to real business problems. Occasional Cypher.