Technical Analysis and Case Studies of Knowledge Graphs by Neo4j

The session, led by Neo4j’s Europe Pre‑Sales and Technical Director Dr. Jesus Barrasa, begins by asking where knowledge lives in a data architecture and uses a financial‑industry example to illustrate that knowledge is embedded in application logic such as SQL scripts and business rules rather than in raw data storage.

It then argues that this implicit knowledge is hard to maintain and proposes knowledge graphs as a way to make data “smarter” by explicitly modeling business‑data relationships as graph edges, adding semantics on top of the business ontology, and off‑loading complex logic from downstream applications.

Next, the talk shows how knowledge graphs enable skill discovery: by ingesting employee, behavior, and product data from multiple sources, building a graph that captures skill taxonomies, and using graph algorithms to identify skill gaps and recommend suitable personnel, as demonstrated in Neo4j’s collaboration with Daimler.

The third part covers metadata management: a knowledge graph can represent not only data assets but also their lineage, source platforms, transformation pipelines, and usage metrics, allowing teams to trace data provenance, troubleshoot pipeline failures, and answer complex queries with Cypher.

The fourth segment discusses semantic search: by extracting entities from documents (e.g., NASA’s technical reports) and linking them in a graph enriched with domain knowledge, users can perform semantic walks to retrieve context‑aware results, far surpassing traditional keyword matching.

Finally, the presenter compares GraphQL and Cypher, noting that GraphQL is suitable for decoupled data access APIs, while Cypher is better for deep graph traversal and complex queries; both can be used together, but Cypher remains the preferred language for graph‑native operations.

A brief Q&A follows, and the session ends with a thank‑you and information about free resource downloads and upcoming events.