Network Analysis
By Seme Research Team · Updated May 22, 2026
Definition
Network Analysis is the study of relationships and connections between entities to understand social structures, information flow, and influence patterns. In identity investigation, network analysis maps professional relationships, organizational affiliations, and social connections using graph theory metrics. Key metrics include: degree centrality (number of direct connections), betweenness centrality (how often an entity lies on the shortest path between others), clustering coefficient (how interconnected an entity's contacts are), and eigenvector centrality (connection quality based on the importance of connected entities). Network analysis reveals hidden structures that are not visible from examining individual profiles — for example, a person with moderate direct connections but high betweenness centrality may be a critical information broker.
How It Works
Network analysis in identity investigation operates on the knowledge graph. First, the graph is constructed from investigation data with entities as nodes and relationships as edges. Second, centrality metrics are calculated for each node to identify key players. Third, community detection algorithms (Louvain, Label Propagation) identify clusters of tightly connected entities — these often represent teams, organizations, or social circles. Fourth, influence propagation models estimate how information or influence flows through the network. Fifth, anomaly detection identifies unusual patterns — entities that bridge otherwise disconnected communities, sudden changes in connection patterns, or unusually dense subgraphs that may indicate coordinated activity.
Example
Analyzing the professional network of a venture capital firm reveals: Partner A has the highest betweenness centrality (bridging the firm's portfolio companies), Partner B has the highest degree centrality (most direct connections), and Analyst C has a surprisingly high eigenvector centrality (connected to highly influential people despite a smaller network). Community detection identifies three clusters: Silicon Valley tech founders, East Coast finance professionals, and academic researchers — with Partner A being the only person bridging all three communities.
Applications
- •Organizational influence mapping for corporate investigations
- •Criminal network analysis in law enforcement
- •Investment network due diligence for financial institutions
- •Academic collaboration pattern analysis for research institutions
Key Statistics
| Metric | Value | Source |
|---|---|---|
| Centrality Metrics | 4 types | Graph theory |
| Community Detection Algorithms | 5+ algorithms | NetworkX / Neo4j |
| Average Network Size | 100-500 nodes | Seme platform data |
| Relationship Types Mapped | 10+ | Seme schema |