Modelling the spillover from online engagement to offline protest: stochastic dynamics and mean-field approximations on networks
Moyi Tian ∗, P. Jeffrey Brantingham, Nancy Rodríguez · 2025
Read Modelling the spillover from online engagement to offline protest: stochastic dynamics and mean-field approximations on networks by Moyi Tian ∗, P. Jeffrey Brantingham, Nancy Rodríguez (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2507-13310.txt.
Introduction
Moyi Tian, Jeffrey Brantingham, and Nancy Rodríguez model how online buzz spills into street protest, in a 2025 paper. Social media now shapes acts from dinner picks to uprisings. Their framework couples an online network, where a topic spreads, to an offline layer where people actually gather. From the full random model they derive simpler average-field versions of several depths, which forecast when surges will hit and how big the spark must be.
One dial rules the spill: the rate at which online heat becomes offline action. Too low, nothing ignites; too high, the heat also fails to build, so outbursts need the middle range. On made-up networks, sparse webs need the deeper approximations, while dense webs are fine with simple ones. On two real networks, added depth did not help. The authors offer the models as planning tools for crowd dynamics.
Catalog record
| CREATOR | Moyi Tian ∗, P. Jeffrey Brantingham, Nancy Rodríguez |
|---|---|
| YEAR | 2025 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 92 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2507-13310
Browse the full library · About the Tower of Babel
Full text
The opening of the work, from the library's hosted text:
Modelling the spillover from online engagement to offline protest: stochastic dynamics and mean-field approximations on networks Moyi Tian ∗, P. Jeffrey Brantingham, Nancy Rodríguez Abstract Social media is transforming various aspects of offline life, from everyday decisions such as dining choices to the progression of conflicts. In this study, we propose a coupled modelling framework with an online social network layer to analyse how engagement on a specific topic spills over into offline protest activities. We develop a stochastic model and derive several mean-field models of varying complexity. These models allow us to estimate the reproductive number and anticipate when surges in activity are likely to occur. A key factor is the transmission rate between the online and offline domains; for offline outbursts to emerge, this rate must fall within a critical range, neither too low nor too high. Additionally, using synthetic networks, we examine how network structure influences the accuracy of these approximations. Our findings indicate that low-density networks need more complex approximations, whereas simpler models can effectively represent higher-density networks. When tested on two real-world networks, however, increased complexity did not enhance accuracy. Keywords: networks, spillover effects, multilayer dynamics, protest modelling, critical transitions, mean-field approximation 2000 Math Subject Classification: 91D30, 37N25, 92D30, 34C60 1 Introduction The use of social media has expanded globally, reaching approximately two-thirds of the global population, and providing a significant segment of our population access to diverse ideas and opinions from individuals worldwide [ 37 , 24 ] . It is fair to say that social media has changed the world, from its effect on mental health to increased misinformation; its effects cannot be understated. Social media has also transformed how, when, and where conflicts occur [ 32 ] . A report by Mercy Corps [ 32 ] studies the effect of social media on spreading ethnic and sectarian tension. This study identifies three primary risks: the amplification of conflict drivers, the targeting of peace advocates, and the spillover of online conflict into the offline world. Of particular relevance to our work is the last, where digital tensions escalate into tangible outcomes such as protests and violence. A natural question arises, “How is social media activity affecting uprisings?” In [ 4 ] , the authors study the Granger causality between social media streams and onsite developments at the Indignados, Occupy, and Brazilian Vinegar protests. The authors found that contentious communication in social media predicted protest activity during the Indignados and Occupy movements. Moreover, they also found bidirectional Granger causality between online and onsite protests in the Occupy series. The role of social media has also been explored in the context of Ukraine’s Euromaidan protests [ 31 ] and the 2011 Egyptian revolution [ 25 ] . In recent years, several studies have developed mathematical frameworks to analyse the spatiotemporal dynamics of protest activities. In 1987, Burbeck et al. applied concepts from epidemiology to model rioting activity, though their model did not consider spatial spread [ 11 ] . This approach initially did not receive much attention, but gained traction in 2017 when Bonnasse-Gahot et al. applied it to model the spatiotemporal dynamics of the 2005 French riots, demonstrating the effectiveness of a non-local version of Burbeck’s epidemiological framework in fitting empirical data [ 7 ] . In addition to the epidemiologically inspired models, a range of other modelling approaches have been developed to capture the dynamics of uprisings, incorporating different mechanisms and sources of influence. Berestycki and team [ 5 ] introduced a model examining the interplay between the evolution of protest activity and social tension; this model was further analysed in [ 6 ] . As another example, Davies et al. in [ 18 ] constructed a mathematical framework to analyse the London riots of 2011 and the corresponding policing strategies. This model incorporated factors such as the infectious nature of participation, distances traveled to riot locations, and the deterrent influence of law enforcement. While these earlier models provided important insights into the mechanisms of protest dynamics, they largely overlooked the growing influence of online social activities in shaping real-world events. Given that over 60 \% of the global population now uses social media [ 37 , 24 ] , that social media connections enhance exposure to offline events, and that the spread of information and tension through social media can also lead to activities in the physical world (offline activity), it is crucial to incorporate online dynamics into models of uprisings and vice versa. To address this, we extend the modelling framework proposed by Burbeck et al. by developing and evaluating a multilayer model that captures the interplay between the spread of tension in online social networks and the dynamics of social uprisings. In this framework, one layer represents the online network that governs information diffusion (tension), while the second layer models offline activities. Our approach builds upon recent developments in other studies, such as [ 30 , 21 ] . In particular, we construct a model that captures the coupled dynamics of online engagement with a contentious issue and corresponding offline protest using a compartmental framework [ 9 ] . Individuals transition between joint online-offline states, with transitions influenced by both social network structure and feedback between layers. The underlying online social network influences transitions from disengaged to engaged states, while only a fraction of engaged individuals take to the streets, protesting for a limited time. We also incorporate feedback: Increased offline protest activity drives further online engagement. We represent online-to-offline spillovers as a Markov process on the multilayer system and develop a range of mean-field approximations that reflect network structure to varying degrees. In simulations on both homogeneous and heterogeneous synthetic networks, we find that higher-order approximations, which capture edge-level dynamics, significantly improve performance in low-density networks but provide limited advantage in denser ones. This trend is less evident in empirical networks, suggesting that real-world structural features not included in our models may play a critical role in shaping these dynamics. To better understand the conditions that drive offline protest outbreaks, we use the continuum models derived by these mean-field approximations and analyse the reproductive numbers, which quantify the potential for activity to spread. Our analysis highlights the importance of the transmission rate between online engagement and offline action: If this rate is too high, the reservoir of protest-susceptible individuals is quickly exhausted; if too low, online flare-ups fail to translate into offline mobilization. A critical range of this transmission rate is therefore essential for substantial offline activity. Outline : Section 2 introduces our model and the motivation behind its development. In Section 3 , we explore the mean-field approximation under the fully-mixed assumption. Section 4 presents two approximation methods for k -regular networks and compares their performance to the stochastic model. An analogous analysis for heterogeneous networks is provided in Section 5 . We conclude with a discussion in Section 6 . 2 Model motivation and background In this section, we introduce our two-layer model and explain its motivation. Section 2.1 provides background context. Section 2.2 presents the overall structure of the model, with Section 2.2.1 focusing on the online layer, where information and tension spread; Section 2.2.2 covering the offline layer, where actions such as boycotts or uprisings