Apply Built-in Knowledge Graph Algorithms in Neo4J + Memgraph

Once we have data points saved in a Knowledge Graph, then we are ready to explore the use of graph algorithms to uncover hidden patterns and high-value insights, integrated with LLM. In this writing, I will use built-in graph data science algorithms in Neo4J and Memgraph — the specific category is called graph community detection — Yep, you may have heard this term from Microsoft GraphRAG technical reading. However, we are going to explore the pure graph algorithm way for now rather than an LLM approach.

What’s New?

Knowledge Graph Data Science / Graph ML is not new —it’s has been widely used in as an analytic tool such as :

• Social network Analysis
• Fraud Detection and Investigation
• IT operation management
• Data Lineage detection
• etc… more.

From my view — what’s worth to visit again is these graph algorithms now been nicely packaged and baked into graph database products as easy-to-use functions or modules. So, this is what drive me to revisit them — exploring how to use these hidden gems — all-in-all help improve analytic workflow and time saving.

Of course, writing our own custom one from scratch as well and a sneak peek on how to use LLM to assist us built cypher queries.

Objectives

- How to call built-in graph algorithms in Neo4j and…