An Agent-Based Model of Posting Behavior During Times of Societal Unrest

Conference paper
Part of the Lecture Notes in Computer Science book series (LNCS, volume 10354)


Social media is increasingly monitored during periods of societal unrest to gauge public response and estimate the duration and severity of related protest events. To this end, we build an agent-based simulation model that accurately describes the shift in posting behavior of users as related to a real historical event. First we define an appropriate indication that an agent has become an “activist”, or someone who disseminates protest-related posts during times of unrest. We then build an agent-based model based on parameters estimated from before and during the protest. We validate our model using a complete collection of Tumblr data from six months prior to the Ferguson protest of 2014, until the state of emergency was lifted. Validation is performed by visual inspection of the similarity of simulated distributions of established emergent metrics to the empirically observed data. Our results show that our model has potential for predicting posting behavior during future protests.


Agent based model Information diffusion Social media Political unrest Tumblr 


  1. 1.
    Bruch, E., Atwell, J.: Agent-based models in empirical social research. Sociol. Methods Res. 44(2), 186–221 (2015)Google Scholar
  2. 2.
    Chung, F., Lu, L.: The average distances in random graphs with given expected degrees. Proc. Natl. Acad. Sci. 99(25), 15879–15882 (2002)MathSciNetCrossRefzbMATHGoogle Scholar
  3. 3.
    Hodas, N.O.: How limited visibility and divided attention constrain social contagion. In: SocialCom, Citeseer (2012)Google Scholar
  4. 4.
  5. 5.
    Weng, L., et al.: Competition among memes in a world with limited attention. Sci. Rep. 2, 1–8 (2012). doi: 10.1038/srep00335

Copyright information

© Springer International Publishing AG 2017

Authors and Affiliations

  1. 1.Indiana UniversityBloomingtonUSA
  2. 2.HRL Laboratories, LLCMalibuUSA

Personalised recommendations