Case Study: Caterpillar improves equipment repair insights with Neo4j

Caterpillar uncovers insights across 27 million documents with Neo4j

Caterpillar faced the challenge of extracting valuable insights from over 27 million disparate equipment repair and maintenance documents. They wanted to build a natural language processing (NLP) tool to uncover hidden connections and trends, creating an open-ended dialog system that could answer any domain-specific question. To achieve this, they turned to the graph database vendor Neo4j.

Using Neo4j, Caterpillar built a solution that employed graph data structures to create a logical form of knowledge. Their architecture used an open-source NLP toolkit to ingest text and the Stanford Dependency Parser to parse it, with Neo4j serving as the core database to find patterns, build hierarchies, and add ontologies. This allowed users to conduct meaningful searches with simple Cypher queries, transforming their massive repository of documents into a powerful knowledge graph for more efficient equipment repair.

Caterpillar