Pith. sign in

REVIEW 3 major objections 4 minor 1 cited by

Spatiotemporal Activity-Driven Networks

T0 review · 3 major / 4 minor · reviewed 2026-08-03 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗ Review JSON

Pith's one-line read Adding a spatial distance kernel to the standard activity-driven temporal-network model makes space act as an implicit memory, spontaneously producing heterogeneous tie weights, clustering, and heavy triangles in the aggregated network, and

desk verdict Solid spatial-temporal network model, but the key social-distancing claim is inconsistent with the model's own contact-generation rule. read the letter →

arxiv 2511.15533 v1 pith:H7X4GGHW submitted 2025-11-19 physics.soc-ph

Zsófia Simon , Jari Saramäki This is my paper · ORCID
classification physics.soc-ph
keywords temporalnetworksactivity-drivenmodelspatialstrongandweaktiesclusteringcoefficientSIRepidemicsocialdistancingrandomgeometricgraphs
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

This paper introduces a spatial variant of the activity-driven temporal-network model in which contacts are chosen preferentially from nearby nodes rather than uniformly at random. Its aim is to show that the spatial rule alone—without any explicit memory or reinforcement mechanism—creates the aggregated-network signatures of human contact: a broad but thin-tailed distribution of link weights, strong and weak ties, high clustering, and triangles heavier than the median link. The authors derive analytic expressions for the ego-net link-weight distribution and for expected triangle weights, and verify them with simulations at N=10^3. They then demonstrate that spatiality changes SIR spreading qualitatively: infection stays localised and spreads more slowly, and a socially-distance-like two-step reduction of the interaction radius that removes only about 1.5% of contacts flattens the epidemic curve, whereas the same reduction applied randomly to a non-spatial network has barely any effect. A sympathetic reader would care because the model is a rare solvable benchmark for spatiotemporal contact structure and gives a principled, minimal handle on why spatial interventions can outperform random contact reduction.

What carries the argument

The central object is the spatial activity-driven model: a discrete-time temporal network in which each node has an activity potential and a fixed position on a torus, and a contact between active node i and candidate j is accepted with probability p_ij = (1 − d_ij/R)/Z_i when d_ij ≤ R, where Z_i normalises the kernel over neighbours inside the disc. The derivation replaces each Z_i with its mean-field value E[Z_i] = π N R²/3, which turns the expected link weight into a linear decline with distance and the expected triangle weight into a linear function of node activities and pairwise distances. Space produces memory because close pairs interact repeatedly, generating heterogeneous weights,

What would settle it

Run the model with a deliberately non-uniform spatial density, e.g., nodes concentrated in two or three clusters, at N=10^3, R=0.2, T=2000, and compare the empirical ego-net link-weight histogram to Eq. 16; substantial deviations in the linear slope or the cutoff at w_max = 3T(a_i+a_j)m/(π N R²) would refute the uniform-density mean-field approximation on which the analytics rest.

Watch

Extended reading notes

Core claim

Central claim: space acts as an inherent memory dimension in temporal networks. Nodes hold fixed torus positions; active nodes make m contacts within radius R with probability proportional to 1−d/R, so expected link weight decays linearly with distance. Aggregated networks gain a broad, thin-tailed weight distribution with strong and weak ties, and triangles of close, active nodes carry above-median weight. Clustering converges to 1−3√3/(4π)≈0.5865, independent of parameters. SIR spreading is slower and localised; shrinking R to remove about 1.5% of contacts cuts the epidemic peak sharply, unlike equal random removal in the non-spatial model. The triangle distribution is a semi-analytic Gaus

Load-bearing premise

The load-bearing premise is that every node sees the same uniform neighbourhood density inside its cutoff radius, so each node's normalisation constant Z_i can be replaced by the global mean π N R²/3; where local density is heterogeneous, the analytic link- and triangle-weight formulas lose their accuracy.

Editorial extensions

If this is right

  • The aggregated network of the spatial activity-driven model has an exponential-like link-weight distribution with both strong and weak ties, unlike the near-uniform weak ties of the original activity-driven model; link weights depend on the aggregation window T, matching empirical observations.
  • In the long-time limit, link density saturates near πR² and the average clustering coefficient converges to 1−3√3/(4π)≈0.5865, independent of R and m, implying a geometric ceiling on clustering in spatially embedded contact networks.
  • SIR dynamics on the spatial model show a delayed and lower infection peak compared with the non-spatial model at matched contact counts, because the infection front is localised and a larger fraction of contacts fail to transmit.
  • A social-distancing intervention modelled by shrinking R in two steps removes only about 1.5% of contacts but markedly flattens and lowers the epidemic curve, whereas random removal of the same number of contacts in the non-spatial model has almost no effect.
  • Because the embedding space need not be geographical, the same framework applies to latent trait spaces, where R defines a social circle and distances encode homophily.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

  • The paper leaves implicit that the uniform-density mean-field assumption is the main bottleneck: a direct test would compare the predicted piecewise-linear link-weight distribution against a network with strongly clustered spatial positions, where local Z_i values vary widely and Eq. 15 should fail.
  • A natural empirical inversion the authors do not perform: measure final link density and clustering in a real contact dataset, invert them to estimate R from πR² and from 0.5865, and check whether the two estimates agree with the observed link-weight distribution.
  • The 'space as memory' interpretation suggests a rewiring experiment: randomise the temporal order of spatial contacts while preserving distances, and the repetition-driven strong ties, heavy triangles, and the clustering ceiling should weaken or disappear, separating spatial memory from pure geometry.
  • The two-step radius reduction is a special case of time-varying R; the framework could be extended to adaptive radius policies where R(t) responds to local infection levels, which the paper does not explore.
Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Referee Report

3 major / 4 minor

Summary. The paper introduces a spatiotemporal extension of the activity-driven temporal network model. Nodes are placed uniformly on a 2D torus; at each discrete time step, active nodes initiate m contacts to neighbours within a cutoff radius R, with probability linearly decreasing with distance (Eq. 2). The authors derive approximate closed-form expressions for the ego-net link-weight distribution (Eqs. 13/16) and for expected triangle weights (Eq. 21), and they report that the model reproduces several stylized facts of social networks: heterogeneous link weights, high clustering converging to the random-geometric-graph value, and triangle weights above the median. They then simulate SIR spreading and claim that shrinking R (social distancing) reduces the total number of contacts by only 1.5% yet dramatically slows the outbreak, in contrast to the same fractional random-removal in the non-spatial activity-driven model, which has little effect.

Significance. If the results are correct, the model is a welcome addition to the small set of analytically tractable temporal-network models with spatial structure. The closed-form clustering limit (Eq. 31) is a strong point, and the link-weight formula provides a clear interpretation of how spatial locality induces heterogeneous tie strengths. The social-distancing application is timely, but its current support is compromised by an inconsistency in the reported contact-count reduction (see Major Comment 1). The paper is clearly written and the derivations are easy to follow; the authors are transparent about the mean-field approximation, although its domain of validity is not explored. The triangle-weight 'analytic' distribution is only semi-empirical. Overall, the structural part of the paper is promising and likely useful to the community, but the intervention claim needs correction before publication.

major comments (3)
  1. [Fig. 10 caption; §II Model (Eq. 2)] The claimed 1.5% reduction in total temporal contacts under the radius intervention is inconsistent with the model definition. In §II, each active node initiates exactly m contacts; Eq. (2) is normalized so that Σ_j p_ij=1, and the cutoff only prevents selection of nodes outside R. For N=10^3 and R=0.075, the expected number of neighbours is NπR^2 ≈ 17.7, so the probability that a node has fewer than m=3 neighbours is ~4×10^-6. Thus shrinking R from 0.15 to 0.075 cannot reduce the total number of contacts by 1.5%; the reduction is less than 0.001%. The comparison with 1.5% random contact removal in the non-spatial model is therefore not an equal-contact-reduction comparison; the spatial intervention is effectively a rewiring rather than a reduction in contact count. Please verify the simulation implementation or report the correct reduction, and re-benchmark the social-distancing claim a
  2. [§III.A, Eqs. (15)-(16) and (20)-(21)] The analytical predictions replace the local normalization constant Z_i with its global mean E[Z_i]=π N R^2/3, assuming uniform density around every node. This is an uncontrolled mean-field approximation. The agreement with simulations is shown only for N=10^3 and R=0.2 (or R=0.15 in later figures). In sparse regimes (small R or low N) the approximation will deteriorate. Since the paper's central claim is analytical tractability, please either provide a bound on the error, test additional parameter values (e.g., R=0.05, N=100), or explicitly restrict the claim of tractability to the dense, uniform case.
  3. [§III.A, Fig. 3 and Eq. (21)] The triangle-weight 'analytic' distribution is not a closed-form prediction. It is obtained by evaluating Eq. (21) for all triangles of the simulated network and then smoothing the resulting values with a Gaussian kernel density estimate, as stated in the text. Thus the distribution itself is not derived analytically; only the expected weight of a triangle given node positions and activities is. The abstract's claim of an analytical result for 'triangles having weights above the median' is therefore not fully supported. Please either derive the distribution (or an approximation to it) or describe this as a numerical validation rather than an analytic one.
minor comments (4)
  1. [References] Reference [10] is incomplete: author names are missing.
  2. [Fig. 9 caption] The caption uses δ=0.9, but the transmission probability is denoted β elsewhere in the text.
  3. [Discussion] The phrase 'space acts as memory' is an interpretation; the model's distance kernel directly encodes a preference for repeated short-range contacts, so the emergence of strong ties is partly built in by construction. Consider softening claims of 'emergence' or explicitly distinguishing mechanism from consequence.
  4. [Eq. (16)] The statement that the analytical result is supported 'up to the highest observed link weight' would benefit from a precise description of how the comparison is made (e.g., binned histogram vs. CDF).

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the analytic derivations are self-contained and not fitted to the simulation outputs; flagged caveats are analyticity/consistency issues, not circular reductions.

full rationale

The paper's central derivation chain is not circular. The link-weight distribution (Eqs. 9-16) follows from the model kernel p_ij (Eq. 2) under an explicitly stated uniform-density approximation E[Z_i] = pi N R^2 / 3; the parameters (T, m, R, activity potentials) are model inputs, and the theoretical curve is compared with, not fitted from, simulation. The triangle-weight formula (Eqs. 18-22) is again obtained by linearity of expectation. The KDE step in the Fig. 3 caption uses the simulated triangle list, so that particular 'analytic' distribution is semi-empirical, but this weakens the analyticity claim rather than making the central result equivalent to its inputs. The asymptotic clustering value in the Appendix is derived independently and agrees with the external RGG result [31]. Self-citations such as [25], [28], and [33] are contextual empirical comparisons and are not load-bearing. Two non-circular concerns should be flagged. First, the claimed 1.5% contact reduction from shrinking R (Fig. 10 caption, Section III C) is difficult to reconcile with Eq. (2): with normalized p_ij, an active node with at least m neighbours still emits m contacts per step, so the stated 1.5% reduction is an internal consistency/correctness issue rather than circularity. Second, the triangle-weight 'prediction' is conditional on the simulated network's triangle set, so it mainly validates the edge-weight formula rather than predicting triangle existence. Neither issue makes a claimed prediction equal to an input by construction.

Assumptions & free parameters 6 free parameters · 5 assumptions · 0 invented entities

No new physical entities are introduced. The model's free parameters (m, R, T, activity exponent, SIR parameters) are tunable inputs or chosen hand parameters; they are not fitted to empirical data. The main hidden assumptions are the uniform-density mean-field replacement of Z_i and the independence-across-time approximation, both of which are acknowledged but not rigorously controlled.

free parameters (6)
  • m = 3 in main simulations (varied in Fig. 5)
    Number of contact attempts per active node per time step; chosen by hand, not estimated from data.
  • R = 0.15-0.2 in main simulations; 0.075-0.1 for social distancing
    Interaction radius on the unit torus; controls the spatial scale of contacts; chosen by hand.
  • T = 1000-2000 time steps (and up to 5000 in Fig. 5)
    Aggregation window; chosen by hand; sets the link-weight scale.
  • Activity power-law exponent = 10
    Shape parameter for the heterogeneous activity distribution; chosen to produce a few highly active nodes, not fitted to data.
  • SIR transmission probability beta = 0.9
    Disease transmission probability in SIR simulations; set well above the epidemic threshold (stated by the authors).
  • SIR recovery time t_r = 3
    Constant recovery time in SIR simulations; chosen by hand.
assumptions (5)
  • domain assumption Node positions are independent and uniformly distributed in a 2D unit torus.
    Used in Eqs. 6-7 for the spatial density of nodes and in the derivation of the analytic link-weight distribution.
  • domain assumption Contact formation is independent across time steps and across nodes.
    Used to write the expected accumulated weight as T*(a_i+a_j)*m*p_ij (Eq. 5); ignores correlations from sampling without replacement and from the activity process.
  • ad hoc to paper Local node density is uniform within the cutoff radius R, so Z_i can be replaced by E[Z_i] = pi N R^2 / 3.
    Mean-field approximation central to Eqs. 13, 16, and 21; only tested at N=10^3, R=0.15-0.2.
  • domain assumption In the T to infinity limit, all pairs within distance R eventually connect, so the aggregated network becomes a random geometric graph.
    Used for d_max approximately pi R^2 and for c_max = 1 - 3*sqrt(3)/(4*pi) (Eqs. 24, 31-32); holds if all activity potentials are strictly positive.
  • domain assumption Activity potentials are drawn from a power-law distribution and normalized by the maximum value.
    Model setup that creates heterogeneous activity; the exponent 10 is chosen, not empirically fitted.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Spatiotemporal Activity-Driven Networks." pith.science (2026). https://pith.science/paper/H7X4GGHW

@misc{pith2026251115533,
  author       = {Pith},
  title        = {Pith review of: Spatiotemporal Activity-Driven Networks},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/H7X4GGHW}},
  note         = {Machine review of arXiv:2511.15533}
}
read the original abstract

Temporal-network models have provided key insights into how time-varying connectivity shapes dynamical processes such as spreading. Among them, the activity-driven model is a widely used, analytically tractable benchmark. Yet many temporal networks, such as those of physical proximity, are also embedded in space, and spatial constraints are known to affect dynamics unfolding on the networks strongly. Despite this, there is a lack of similar simple and solvable models for spatiotemporal contact structures. Here, we introduce a spatial activity-driven model in which short-range contacts are more frequent. This model is analytically tractable and captures the joint effects of space and time. We show analytically and numerically that the model reproduces several characteristic features of social and contact networks, including strong and weak ties, clustering, and triangles having weights above the median. These traits can be attributed to space acting as a form of memory. Simulations of spreading dynamics on top of the model networks further illustrate the role of space, highlighting how localisation slows down spreading. Furthermore, the framework is well-suited for modelling social distancing in a principled way as an intervention measure aimed at reducing long-range links. We find that, unlike for non-spatial networks, even a small spatially targeted reduction in the total number of contacts can be very effective. More broadly, by offering a tractable framework, the model enables systematic exploration of dynamical processes on spatiotemporal networks.

Figures

Figures reproduced from arXiv: 2511.15533 by the authors.

Figure 1
Figure 1. FIG. 1: Visualisation of the spatially preferential contact process. [PITH_FULL_IMAGE:figures/full_fig_p003_1.png] view at source ↗
Figure 2
Figure 2. FIG. 2: Comparison of theoretical [PITH_FULL_IMAGE:figures/full_fig_p004_2.png] view at source ↗
Figure 3
Figure 3. FIG. 3: Analytical approximation [PITH_FULL_IMAGE:figures/full_fig_p005_3.png] view at source ↗
Figures from the paper (7 more)
Figure 4
Figure 4. Figure 4: FIG. 4: Correlation (linear regression) results for spatiality and activity of triangles with triangle weight. Blue dots [PITH_FULL_IMAGE:figures/full_fig_p006_4.png]
Figure 5
Figure 5. Figure 5: FIG. 5: Evolution of link density for different [PITH_FULL_IMAGE:figures/full_fig_p006_5.png]
Figure 6
Figure 6. Figure 6: FIG. 6: Comparison of empirical link weight [PITH_FULL_IMAGE:figures/full_fig_p007_6.png]
Figure 8
Figure 8. Figure 8: FIG. 8: Comparison of the evolution of the average [PITH_FULL_IMAGE:figures/full_fig_p008_8.png]
Figure 9
Figure 9. Figure 9: FIG. 9: SIR spreading process on top of a spatiotemporal network with [PITH_FULL_IMAGE:figures/full_fig_p009_9.png]
Figure 10
Figure 10. Figure 10: FIG. 10: Mean simulated SIR infection curves with zoomed-in inset (panel A) and cumulative mean curves for the [PITH_FULL_IMAGE:figures/full_fig_p010_10.png]
Figure 11
Figure 11. Figure 11: FIG. 11: The evolution of mean failed infection probabilities [PITH_FULL_IMAGE:figures/full_fig_p010_11.png]

Discussion (0). Continue with ORCID to comment.

Forward citations

Cited by 1 Pith paper

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score. Full citation record

  1. Streaming Model Cascades for Semantic SQL

    cs.DB 2026-04 unverdicted novelty 5.5 of 10

    SUPG-IT and GAMCAL route streaming semantic-SQL rows through cheap proxies with joint precision/recall guarantees or a single cost-error tradeoff, cutting oracle calls while keeping high F1.

Reference graph

Works this paper leans on

41 extracted references · cited by 1 Pith paper

  1. [1]

    Temporal networks,

    P. Holme and J. Saramäki, “Temporal networks,”Physics Reports, vol. 519, no. 3, pp. 97–125, 2012

    2012

  2. [2]

    Then, each node may or may not become active according to its own assigned activity potential

  3. [3]

    Each active node can initiatemcontacts to be formed withmother nodes, which need not be ac- tive at the time. Themnodes get selected accord- ing to a distance kernel so that for active nodei, potential neighboursjare assigned a contact prob- ability inversely proportional to their distance to the active node, i.e., tod(i, j), wheredis the 2D toroidal dist...

  4. [4]

    The spatial location and activity potential of nodes do not change

    At the next time stept′ =t+ ∆t(for simplicity, we use∆t= 1), the process returns to step 1, re- verting the network toNdisconnected nodes. The spatial location and activity potential of nodes do not change

  5. [5]

    The weightw ij of a link is defined as the total number of contacts between the nodes iandj

    The aggregated networkGat timeTis the union of all networksG t up to that time, i.e.G=S t∈{0,...,T} Gt. The weightw ij of a link is defined as the total number of contacts between the nodes iandj. For simplicity, we use a linear kernel for the contact probability with a cut-off radiusRsuch that ifd(i, j)> R, wheredis the 2D toroidal distance, the probabil...

  6. [6]

    Masuda and P

    N. Masuda and P. Holme, Eds.,Temporal Network Epi- demiology. Singapore: Springer, 2017

    2017

  7. [7]

    Saramäki and P

    J. Saramäki and P. Holme", Eds.,Temporal Network Theory, 2nd edition. Switzerland: Springer, 2023

    2023

  8. [8]

    Impact of non-poissonian activity patterns on spreading processes,

    A. Vazquez, B. Rácz, A. Lukács, and A.-L. Barabási, “Impact of non-poissonian activity patterns on spreading processes,”Physical Review Letters, vol. 98, p. 158702, 2007

    2007

Show all 41 references
  1. [9]

    Small but slow world: How network topology and burstiness slow down spread- ing,

    M. Karsai, M. Kivelä, R. Pan, K. Kaski, J. Kertész, A. Barabási, and J. Saramäki, “Small but slow world: How network topology and burstiness slow down spread- ing,”Physical Review E, vol. 83, p. 025102, 2011

    2011

  2. [10]

    Dynamics of cascades on burstiness-controlled temporal networks,

    S. Unicomb, G. Iñiguez, J. P. Gleeson, and M. Karsai, “Dynamics of cascades on burstiness-controlled temporal networks,”Nature Communications, vol.12, no.1, p.133, 2021

    2021

  3. [11]

    Activity driven modeling of time vary- ing networks,

    N. Perra, B. Gonçalves, R. Pastor-Satorras, and A. Vespignani, “Activity driven modeling of time vary- ing networks,”Scientific Reports, vol. 2, no. 1, p. 469, 2012

    2012

  4. [12]

    Epidemic spreading on activity-driven networks with attractiveness,

    I. Pozzana, K. Sun, and N. Perra, “Epidemic spreading on activity-driven networks with attractiveness,”Physi- cal Review E, vol. 96, no. 4, p. 042310, 2017

    2017

  5. [13]

    Con- trolling contagion processes in activity driven networks,

    S. Liu, N. Perra, M. Karsai, and A. Vespignani, “Con- trolling contagion processes in activity driven networks,” Physical Review Letters, vol. 112, no. 11, p. 118702, 2014

    2014

  6. [14]

    From calls to communities: a model for time-varying so- cial networks,

    “From calls to communities: a model for time-varying so- cial networks,”The European Physical Journal B, vol. 88, p. 301, 2015

    2015

  7. [15]

    Spatial networks,

    M. Barthélemy, “Spatial networks,”Physics Reports, vol. 499, no. 1–3, pp. 1–101, 2011

    2011

  8. [16]

    Spatiotemporal signal propagation in complex networks,

    C. Hens, U. Harush, S. Haber, R. Cohen, and B. Barzel, “Spatiotemporal signal propagation in complex networks,”Nature Physics, vol. 15, pp. 403–412,

  9. [17]

    Interaction data from the copenhagen net- works study,

    P. Sapiezynski, A. Stopczynski, D. D. Lassen, and S. Lehmann, “Interaction data from the copenhagen net- works study,”Scientific Data, vol. 6, no. 1, p. 315, 2019

    2019

  10. [18]

    Multiscale mobility networks 13 and the spatial spreading of infectious diseases,

    D. Balcan, V. Colizza, B. Gonçalves, H. Hu, J. J. Ra- masco, and A. Vespignani, “Multiscale mobility networks 13 and the spatial spreading of infectious diseases,”Proceed- ings of the National Academy of Sciences, vol.106, no.51, pp. 21484–21489, 2009

    2009

  11. [19]

    Natural hu- man mobility patterns and spatial spread of infectious diseases,

    V. Belik, T. Geisel, and D. Brockmann, “Natural hu- man mobility patterns and spatial spread of infectious diseases,”Physical Review X, vol. 1, p. 011001, 2011

    2011

  12. [20]

    The hidden geometry of complex, network-driven contagion phenomena,

    D. Brockmann and D. Helbing, “The hidden geometry of complex, network-driven contagion phenomena,”Sci- ence, vol. 342, p. 1337, 2013

    2013

  13. [21]

    Simulation of an SEIR infectious disease model on the dynamic con- tact network of conference attendees

    J. Stehlé, N. Voirin, A. Barratet al., “Simulation of an SEIR infectious disease model on the dynamic con- tact network of conference attendees.”BMC Med., vol. 9, p. 87, 2011

    2011

  14. [22]

    Population flow drives spatio-temporal distribu- tion of COVID-19 in China,

    J. Jia, X. Lu, Y. Yuan, G. Xu, J. Jia, and N. Chris- takis, “Population flow drives spatio-temporal distribu- tion of COVID-19 in China,”Nature, vol. 582, pp. 389– 394, 2020

    2020

  15. [23]

    Covid-19 lockdown induces disease-mitigating structural changes in mobility networks,

    F. Schlosser, B. F. Maier, O. Jack, D. Hinrichs, A. Zachariae, and D. Brockmann, “Covid-19 lockdown induces disease-mitigating structural changes in mobility networks,”Proceedings of the National Academy of Sci- ences, vol. 117, no. 52, pp. 32883–32890, 2020

    2020

  16. [24]

    Comparing large-scale computational approaches to epidemic model- ing: Agent-basedversusstructuredmetapopulationmod- els,

    M. Ajelli, B. Gonçalves, D. Balcan, V. Colizza, H. Hu, J. Ramasco, S. Merler, and A. Vespignani, “Comparing large-scale computational approaches to epidemic model- ing: Agent-basedversusstructuredmetapopulationmod- els,”BMC Infectious Diseases, vol.10, no.1, p.190, 2010

    2010

  17. [25]

    The effect of travel restrictions on the spread of the 2019 novel coronavirus (COVID-19) out- break,

    C. M. others, “The effect of travel restrictions on the spread of the 2019 novel coronavirus (COVID-19) out- break,”Science, vol. 368, pp. 395–400, 2020

    2019

  18. [26]

    Modeling the spatiotemporal epidemic spreading of COVID-19 and the impact of mobility and social distancing interventions,

    A. Arenas, W. Cota, J. Gómez-Gardeñes, S. Gómez, C. Granell, J. Matamalas, D. Soriano-Paños, and B. Steinegger, “Modeling the spatiotemporal epidemic spreading of COVID-19 and the impact of mobility and social distancing interventions,”Physical Review X, vol. 10, p. 041055, 2020

    2020

  19. [27]

    Time as a lim- ited resource: Communication strategy in mobile phone networks,

    G. Miritello, E. Moro, R. Lara, R. Martínez-López, S. G. B. Roberts, and R. I. M. Dunbar, “Time as a lim- ited resource: Communication strategy in mobile phone networks,”Social Networks, vol. 35, pp. 89–95, 2013

    2013

  20. [28]

    Modeling commu- nicable diseases, human mobility, and epidemics: A re- view,

    D. Soriano-Paños, W. Cota, S. Ferreira, G. Ghoshal, A. Arenas, and J. Gómez-Gardeñes, “Modeling commu- nicable diseases, human mobility, and epidemics: A re- view,”Annalen der Physik, vol. 534, p. 2100482, 03 2022

    2022

  21. [29]

    Effect of manual and digital contact trac- ing on covid-19 outbreaks: a study on empirical contact data,

    A. Barrat, C. Cattuto, M. Kivelä, S. Lehmann, and J. Saramäki, “Effect of manual and digital contact trac- ing on covid-19 outbreaks: a study on empirical contact data,”Journal of the Royal Society Interface, vol. 18, no. 178, p. 20201000, 2021

    2021

  22. [30]

    Analysis of a large-scale weighted network of one-to-one human communication,

    J.-P. Onnela, J. Saramäki, J. Hyvönen, G. Szabó, M. A. de Menezes, K. Kaski, A.-L. Barabási, and J. Kertész, “Analysis of a large-scale weighted network of one-to-one human communication,”New Journal of Physics, vol. 9, p. 179, 2007

    2007

  23. [31]

    Orbital prefrontal cortex volume predicts social network size: an imaging study of indi- vidual differences in humans,

    J. Powell, P. A. Lewis, N. Roberts, M. Garcia-Finana, and R. I. M. Dunbar, “Orbital prefrontal cortex volume predicts social network size: an imaging study of indi- vidual differences in humans,”Proceedings of the Royal Society B: Biological Sciences, vol. 279, pp. 2157–2162, 2012

    2012

  24. [32]

    The mathematics of infectious diseases,

    H. W. Hethcote, “The mathematics of infectious diseases,”SIAM Review, vol. 42, pp. 599–653,

  25. [33]

    Effects of time window size and placement on the structure of aggregated networks,

    G. Krings, M. Karsai, S. Bernharsson, V. D. Blondel, and J. Saramäki, “Effects of time window size and placement on the structure of aggregated networks,”EPJ Data Sci- ence, vol. 1, p. 4, 2012

    2012

  26. [34]

    Social network analysis: Methods and applications,

    S. Wasserman, “Social network analysis: Methods and applications,”The Press Syndicate of the University of Cambridge, 1994

    1994

  27. [35]

    Crossover from scale-free to spatial net- works,

    M. Barthélemy, “Crossover from scale-free to spatial net- works,”Europhysics Letters, vol. 63, pp. 915–921, 2003

    2003

  28. [36]

    Random geometric graphs,

    J. Dall and M. Christensen, “Random geometric graphs,” Physical Review E, vol. 66, no. 1, p. 016121, 2002

    2002

  29. [39]

    Structure and tie strengths in mobile communication networks,

    J.-P. Onnela, J. Saramäki, J. Hyvönen, G. Szabó, D. Lazer, K. Kaski, J. Kertész, and A.-L. Barabási, “Structure and tie strengths in mobile communication networks,”Proceedings of the National Academy of Sci- ences, vol. 104, no. 18, pp. 7332–7336, 2007

    2007

  30. [40]

    Geographical dispersal of mobile communication networks,

    R. Lambiotte, V. D. Blondel, C. de Kerchove, E. Huens, C. Prieur, Z. Smoreda, and P. V. Dooren, “Geographical dispersal of mobile communication networks,”Physica A: Statistical Mechanics and its Applications, vol. 387, no. 21, pp. 5317–5325, 2008

    2008

  31. [41]

    Geographic constraints on social network groups,

    J.-P. Onnela, S. Arbesman, M. C. González, A.-L. Barabási, and N. A. Christakis, “Geographic constraints on social network groups,”PLoS ONE, vol. 6, p. e16939, 2011

    2011

  32. [2000]

    Available: https://doi.org/10.1137/ S0036144500371907

    [Online]. Available: https://doi.org/10.1137/ S0036144500371907

  33. [2019]

    Available: https://www.nature.com/ articles/s41567-018-0409-0

    [Online]. Available: https://www.nature.com/ articles/s41567-018-0409-0

Pith tools

Reviewed August 3, 2026 · model on record in the stance chip above.