Connecting Dots That Others Miss
Connecting Dots That Others Miss
Tabular data hides relationships. A spreadsheet shows you that John Smith is connected to Acme Corp, but it cannot show you that John's wife's brother is the CFO of Acme's biggest supplier, who is currently under SEC investigation.
Link analysis surfaces these hidden patterns by treating data as a network of connected entities rather than rows in a table.
Start with what you know: a name, email, company, or event. This is your investigation's anchor point.
Query all sources for entities directly connected to your seed. Employment, ownership, communication, financial ties.
For each first-degree connection, decide whether to expand further. Prioritize high-risk or anomalous nodes.
Look for clusters, bridges, and isolated nodes. Calculate centrality metrics. Note temporal patterns.
Create a clear visual that tells the story. Highlight the key relationships that answer your intelligence question.
Not all connections are equal. When building your graph, classify relationships by type and strength.
| TYPE | EXAMPLES | STRENGTH |
|---|---|---|
| Family | Spouse, parent, sibling, in-law | HIGH |
| Financial | Shared accounts, transactions, beneficial ownership | HIGH |
| Legal | Co-defendant, witness, attorney-client | HIGH |
| Professional | Same employer, board membership, contractor | MEDIUM |
| Digital | Shared IP, device, email domain | MEDIUM |
| Social | Mutual followers, tagged photos, group membership | LOW |
An entity with many connections who appears nowhere in public leadership. Often the real decision-maker operating through proxies. Look for high degree centrality combined with low public visibility.
An entity connecting two otherwise separate clusters. Could indicate a broker, intermediary, or point of vulnerability. Identified by high betweenness centrality.
Multiple entities with minimal activity, all connected to a single hub. Classic structure for money laundering, sanctions evasion, or fraud. Look for entities created around the same time with similar registration patterns.