Temporal Social Network Analysis
Temporal Social Network Analysis (Longitudinal and Time-Varying Network Analysis) · Also known as: TSNA, longitudinal social network analysis, time-varying network analysis, dynamic SNA
Temporal Social Network Analysis (TSNA) extends classic social network analysis by treating networks as time-varying structures. Rather than aggregating all ties into a single static snapshot, TSNA tracks when ties form, persist, and dissolve, enabling researchers to study how social structures evolve and how dynamic connectivity shapes diffusion, influence, and inequality 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.
+4 more
When to use it
Use TSNA when you have longitudinal relational data (panel waves, interaction logs, time-stamped events) and a research question that concerns change: how networks grow or shrink, when actors become influential, whether communities form and dissolve, or how information diffuses along time-respecting paths. It is the appropriate choice when assuming a static network would distort conclusions. Do NOT use TSNA when data are purely cross-sectional with no temporal ordering, when observation windows are too short to observe meaningful change, or when sample sizes per time slice fall too low for reliable metric estimation. In those cases, standard static SNA or ERGM on the aggregated network is more appropriate.
Strengths & limitations
- Captures the sequential and durational structure of ties that static snapshots erase.
- Enables causal-direction reasoning: earlier ties can influence later ones, supporting network process models.
- Temporal centrality metrics reveal when actors are influential, not just if they are influential.
- Compatible with diffusion, epidemic, and social influence models that require time-ordered contact sequences.
- Supports both descriptive trajectory analysis and inferential modelling (temporal ERGM, SAOM).
- Requires time-stamped or panel-wave relational data, which are costly and rare in many disciplines.
- Snapshot granularity is a researcher degree of freedom that can substantially affect conclusions.
- Per-slice networks are often sparse, making centrality estimates unstable and community detection unreliable.
- Computational cost scales with the number of time points and network size; large temporal networks demand specialised tools.
Frequently asked
What data format does temporal SNA require?
You need relational data with a time component: either repeated cross-sectional network surveys (panel waves), or an event log recording when each directed or undirected tie was active (onset, terminus, or discrete timestamps). Actor-level attribute data should be aligned to the same time points.
How do I choose the right snapshot width?
There is no universal rule. The window should be wide enough that each slice contains enough ties for stable metrics, but narrow enough to detect the changes you care about. Always run a sensitivity analysis across two or three alternative granularities and report whether conclusions change.
Is temporal SNA the same as dynamic network analysis?
The terms are often used interchangeably. Some authors reserve 'dynamic' for continuously evolving networks (e.g., agent-based models), while 'temporal' emphasises empirical time-stamped data. In practice, both refer to the study of time-varying network structure, and the methods overlap substantially.
What software supports temporal SNA?
In R, the tsna and networkDynamic packages (developed by the Statnet team) are the standard tools for longitudinal network data. Python's networkx supports basic temporal operations, and the teneto package provides temporal centrality metrics. Gephi supports animated temporal visualisation.
Can I use ERGM-style inference on temporal networks?
Yes. Temporal ERGMs (tERGMs) model the probability of a tie existing at time t given the network state at t-1, allowing inference on tie formation and dissolution processes while controlling for structural endogeneity. Stochastic actor-based models (RSiena) provide an alternative, actor-oriented framework.
Sources
- Holme, P., & Saramäki, J. (2012). Temporal networks. Physics Reports, 519(3), 97–125. DOI: 10.1016/j.physrep.2012.03.001 ↗
- Moody, J., McFarland, D., & Bender-deMoll, S. (2005). Dynamic network visualization. American Journal of Sociology, 110(4), 1206–1241. DOI: 10.1086/421509 ↗
How to cite this page
ScholarGate. (2026, June 3). Temporal Social Network Analysis (Longitudinal and Time-Varying Network Analysis). ScholarGate. https://scholargate.app/en/network-analysis/temporal-social-network-analysis
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.
- Multiplex Network AnalysisNetwork analysis↔ compare
- Network Diffusion AnalysisNetwork analysis↔ compare
- Social Network AnalysisNetwork analysis↔ compare
- Temporal Community DetectionNetwork analysis↔ compare