Temporal Degree Centrality
Temporal Degree Centrality in Time-Varying Networks · Also known as: time-varying degree centrality, dynamic degree centrality, temporal node degree, TDC
Temporal degree centrality extends the classic degree centrality to time-varying networks by counting how many distinct contacts a node accumulates over time. Rather than collapsing a dynamic network into a single static graph, it preserves the temporal order of edges, yielding a more faithful measure of a node's activity and reachability across the observation window.
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 temporal degree centrality when your network data are collected with timestamps — such as email logs, mobile call records, proximity sensor data, or social media interactions — and you want to identify nodes that are persistently active or that accumulate the most contacts over time. It is preferable to static degree centrality whenever event ordering matters for diffusion or influence. Do not use it on cross-sectional network data recorded at a single point in time, or when the temporal resolution of your data is too coarse to distinguish meaningful event sequences.
Strengths & limitations
- Captures sustained activity rather than a single frozen snapshot, giving a more dynamic picture of node importance.
- Sensitive to bursting and periodicity: nodes active across many periods are distinguished from those active in one large burst.
- Computationally straightforward: it extends simple degree counting to a temporal dimension without requiring path enumeration.
- Compatible with standard network metrics: temporal degree rankings correlate predictably with static degree in well-mixed networks, enabling benchmarking.
- Applicable to weighted temporal networks by summing contact durations or interaction intensities instead of binary edge counts.
- Requires time-stamped edge data; inapplicable when only a static or aggregated adjacency matrix is available.
- Ignores path structure and reachability — a node may have high temporal degree but be poorly positioned to relay information if its contacts do not overlap in time.
- Results depend heavily on the chosen time-slice resolution: too coarse and temporal structure is lost; too fine and most slices are empty.
- Does not account for the direction of influence in directed networks without additional in/out-degree decomposition.
Frequently asked
How is temporal degree centrality different from static degree centrality?
Static degree centrality collapses all edges into a single snapshot and counts each unique neighbour once. Temporal degree centrality retains the timestamps and sums or averages instantaneous degree across all time steps, so repeated contacts with the same neighbour are counted each time they occur, reflecting actual interaction volume.
What time-slice resolution should I choose?
There is no universal rule: the resolution should match the natural time scale of your process of interest. For face-to-face contact networks, minutes or hours are typical; for citation or collaboration networks, years may be appropriate. Always perform a sensitivity analysis by computing centrality at several resolutions and checking whether rankings change substantially.
Can temporal degree centrality be used on directed networks?
Yes. In directed temporal networks you decompose into in-degree (contacts received) and out-degree (contacts initiated) at each time step, then aggregate separately. This reveals whether a node is primarily a broadcaster, a receiver, or both.
Is high temporal degree sufficient to identify influential spreaders?
Not necessarily. A node may accumulate many contacts but still be a poor spreader if those contacts are concentrated in a single burst after the epidemic peak, or if its neighbours are not temporally reachable from each other. Pair temporal degree with temporal reachability or time-respecting betweenness for a fuller picture.
What software can compute temporal degree centrality?
The Python libraries NetworkX (with temporal graph extensions), Pathpy, and Teneto all support temporal network analysis including temporal degree. In R, the tsna package provides temporal SNA metrics compatible with the statnet suite.
Sources
- Holme, P. & Saramaki, J. (2012). Temporal networks. Physics Reports, 519(3), 97–125. DOI: 10.1016/j.physrep.2012.03.001 ↗
- Kim, H. & Anderson, R. (2012). Temporal node centrality in complex networks. Physical Review E, 85(2), 026107. DOI: 10.1103/PhysRevE.85.026107 ↗
How to cite this page
ScholarGate. (2026, June 3). Temporal Degree Centrality in Time-Varying Networks. ScholarGate. https://scholargate.app/en/network-analysis/temporal-degree-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.
- Degree CentralityNetwork analysis↔ compare
- Temporal Betweenness CentralityNetwork analysis↔ compare
- Temporal Closeness CentralityNetwork analysis↔ compare
- Temporal Eigenvector CentralityNetwork analysis↔ compare
- Temporal PageRankNetwork analysis↔ compare
- Temporal Social Network AnalysisNetwork analysis↔ compare