Graph Analysis Engine Domain

Overview#

Finding the most connected individual in a money laundering network, or detecting that three seemingly unrelated accounts form a transaction cycle, requires graph mathematics that standard database queries cannot provide efficiently. The Graph Analysis Engine exposes pre-computed PageRank scores, Louvain community clusters, and network motif patterns, making those results available instantly without running live graph traversals on every query.

Key Features#

• PageRank centrality scoring for identifying influential entities in the network
• Louvain community detection for discovering entity clusters and groups
• Network motif pattern identification: triangles, stars, chains, cycles, and cliques
• Graph statistics including node count, edge count, density, and modularity
• Entity search across the graph analysis results
• Pagination support for large result sets
• Periodic data refresh with configurable schedules per metric type

Use Cases#

Relevant sectors include financial crime, law enforcement, and intelligence agencies.

• Identifying the most influential entities in a relationship network using centrality scores
• Discovering hidden communities and clusters within investigative data
• Detecting suspicious network patterns such as transaction cycles or hub-and-spoke structures
• Generating graph-level statistics for dashboard views and analytical reporting

Integration#

The Graph Analysis Engine domain integrates with Graph for visualisation, Network for network analysis, Analytics for dashboard reporting, Entity for entity management, and Relationship for connection data.

Open Standards#

• GraphQL (June 2018 specification): all Graph Analysis Engine queries (top entities, communities, motifs, statistics, and entity search) are exposed exclusively through a Strawberry GraphQL schema, using the GraphQL type system and query execution model as defined in the specification.
• openCypher: the graph database layer uses Cypher, the openCypher-specified declarative graph query language, to traverse and match nodes and relationships stored in Neo4j.
• Bolt Protocol: the Neo4j driver connects to the graph database over the Bolt binary wire protocol (bolt:// and encrypted neo4j+s:// URIs), Neo4j's client, server transport specification.
• ISO/IEC 39075 (GQL): the graph query language work aligns with the ISO/IEC 39075 Graph Query Language standard, of which openCypher is a principal contributing specification.
• RFC 7519 (JSON Web Tokens): every resolver enforces IsAuthenticated, which verifies an RS256-signed JWT against the platform JWKS endpoint before any graph data is returned.
• RFC 8259 (JSON): all GraphQL responses, including paginated entity lists, community summaries, motif catalogues, and graph statistics, are serialised as JSON per the IETF JSON specification.

Last Reviewed: 2026-02-09 Last Updated: 2026-04-14