Clustering Coefficient
Measure how tightly knit your network is
Clustering coefficient quantifies the degree to which nodes cluster together - the prevalence of triangles compared to all possible triangles.
What It Computes
Global metric (0 to 1): Probability that two neighbors of a node are also connected.
When to Use It
- Network health: How cohesive is the network?
- Community strength: Higher = stronger communities
- Compare networks: Benchmark against similar networks
Performance
Time: O(E × avg_degree)
Scales to: 10M edges
Example
Interpretation:
- ~0.0: No clustering (tree-like)
- 0.3-0.6: Typical social networks
- ~1.0: Highly clustered (many triangles)
Use Cases
Social Network Health
Healthy communities have high clustering (0.4+)
Fraud Detection
Fake account networks have LOW clustering (no mutual connections)
Network Evolution
Track clustering over time: