Graph databases store data as entities and relationships.

They are useful when relationships are the primary thing being queried, rather than incidental foreign keys or references. Examples include social connections, permissions, fraud rings, dependency graphs, knowledge graphs, network topology, and supply-chain relationships.

The practical question is not "does my data have relationships?" Most data does. The better question is: "Do my important queries traverse relationships, follow paths, or search for patterns in connected data?"

If the answer is yes, a graph database may fit. If the system mostly retrieves entities by ID, filters rows, or aggregates facts, a relational, document, key-value, or analytical database may be simpler.