Dynamic Closeness Centrality
Dynamic Closeness Centrality in Temporal Networks · Also known as: temporal closeness centrality, time-varying closeness centrality, evolving network closeness, dynamic CC
Dynamic closeness centrality extends classic closeness centrality to temporal networks by computing shortest time-respecting paths — paths that traverse edges in chronological order — and averaging inverse distances across all time windows. It reveals which nodes are most efficiently reached within an evolving network, tracking how a node's centrality rises and falls as connections appear and disappear over time.
Read the full method
Sign in with a free account to read this section.
Method map
The neighbourhood of related methods — select a node to explore.
When to use it
Use dynamic closeness centrality when your network data carries timestamps and the sequence of interactions matters — for example, contact tracing, email or messaging logs, co-authorship over years, or financial transaction networks. It answers questions like 'who was most efficiently connected during a given outbreak phase?' or 'which collaborators bridged communities at the key moment a field shifted?' Do not apply it to cross-sectional, aggregated network snapshots where edge timing is unknown, nor to networks where interactions have no meaningful chronological order. Computationally, it scales worse than static closeness, so very large temporal networks (millions of events) require approximation algorithms or sampling.
Strengths & limitations
- Captures the causal structure of real interactions by respecting the chronological order of edges.
- Reveals temporal patterns — rising and falling influence — invisible to static centrality measures.
- Identifies nodes that are structurally central at critical time points, improving targeting in diffusion or intervention studies.
- Compatible with real-world data formats: timestamped event logs common in digital trace data.
- Reachability-adjusted variants handle disconnected or sparse temporal networks gracefully.
- Computationally expensive: temporal shortest-path algorithms are O(n·m) or worse, limiting scalability to large networks.
- Results are sensitive to the chosen time window granularity — fine-grained or coarse-grained windows can yield substantially different centrality rankings.
- Requires high-quality timestamped data; missing or imprecise timestamps bias temporal distance estimates.
- Interpretation is more complex than static closeness, making communication to non-specialist audiences harder.
Frequently asked
How is dynamic closeness centrality different from static closeness centrality?
Static closeness uses shortest paths in a timeless graph where all edges are simultaneously available. Dynamic closeness restricts paths to those that respect edge timestamps, so only causally valid routes count. This can dramatically change rankings: a node with many simultaneous connections but poor timing can score lower than a node with fewer but better-timed contacts.
What data format do I need?
You need a timestamped edge list — at minimum three columns: source node, target node, and event time. Higher-resolution timestamps generally improve accuracy. Aggregated adjacency matrices without timestamps cannot support this analysis.
How do I choose the time window?
There is no universal rule. Common approaches include using natural breakpoints in the data (e.g., days, weeks, semesters), running sensitivity analyses across multiple window sizes, or using sliding windows with a fixed width. Always report which window was used and check whether key findings hold across alternative window choices.
Can I use this on very large networks?
Exact computation becomes impractical above a few hundred thousand events. For larger networks, approximation algorithms, node sampling, or parallel implementations (available in some graph libraries) are necessary. Report the approximation strategy used.
What software implements dynamic closeness centrality?
The Python library Teneto provides temporal network centrality measures including temporal closeness. The R package tsna offers related temporal SNA tools. Custom implementations using NetworkX (Python) are also common in the literature.
Sources
- Tang, J., Musolesi, M., Mascolo, C., Latora, V. & Nicosia, V. (2010). Analysing information flows and key mediators through temporal centrality metrics. Proceedings of the 3rd Workshop on Social Network Systems (SNS '10). ACM. DOI: 10.1145/1852658.1852661 ↗
- Holme, P. & Saramäki, J. (2012). Temporal networks. Physics Reports, 519(3), 97–125. DOI: 10.1016/j.physrep.2012.03.001 ↗
How to cite this page
ScholarGate. (2026, June 3). Dynamic Closeness Centrality in Temporal Networks. ScholarGate. https://scholargate.app/en/network-analysis/dynamic-closeness-centrality
Which method?
Set this method beside its closest kin and read them side by side — the library lays the books on the table; the choice is yours.
- Betweenness CentralityNetwork analysis↔ compare
- Closeness CentralityNetwork analysis↔ compare
- Dynamic Degree CentralityNetwork analysis↔ compare
- Temporal Community DetectionNetwork analysis↔ compare
- Temporal Social Network AnalysisNetwork analysis↔ compare