Human Dynamics is a transdisciplinary research field for understanding and analyzing dynamic patterns, relationships, and changes in human activities, behaviors, and communication. This IBSS large interdisciplinary project will study human dynamics across social media and social networks and focus on the spatiotemporal modeling of information diffusion and the inter-correlation between online activities and real world human behaviors. Two scenarios, public response to disaster warnings and referendum of controversial social topics at state or national level, will be used to validate and improve a new communication theory, Multilevel Model of Meme Diffusion (M3D). M3D is a new framework designed for describing online communication and the diffusion of memes (social media messages) via different social networks. M3D theory was developed by Spitzberg (Co-PI) in 2012 from our current NSF funded project (“Mapping ideas from cyberspace to realspace”, #1028177, 2010-2014, http://mappingideas.sdsu.edu/).

These two scenarios are selected because the outcomes of human activities in both scenarios (disaster warnings and referendum) are detectable and measurable. The analysis of spatiotemporal correlations of human behaviors and response will be used to validate and calibrate our computational models (agent-based modeling and social network algorithms) and the M3D theory for better understanding human dynamics and the diffusion of online information. This interdisciplinary research project will build a strong research team with experts from social media analytics, social network analysis, geographic information science (GIScience), computational linguistics, agent-based modeling, human behavior analysis, and communication theories, to study human dynamics. This research extends the original interdisciplinary research team from our current NSF project (#1028177) to provide more comprehensive research capabilities by adding new team members, including computational scientists in social network analysis (Jin), high performance computing (Xuan), agent-based modeling (ABM) (Lee and Ye), and human behavioral analysis (Corliss). Our interdisciplinary research team will study spatiotemporal diffusion patterns of human messages, activities, and communications by using both computational methods (social network analysis, geographic information systems (GIS), spatial statistics, machine-learning, and agent-based modeling) and social science approaches (qualitative analysis, inferential statistics, and behavior analysis). The study will help social and behavioral scientists understand the complicated mechanisms of human communications in both cyberspace (online) and the real world (offline).

In our research design, we treat each social media message as a meme from a communication perspective. Memes are replicable messages (Blackmore, 1999; Heylighen & Chielens, 2009; Paull, 2009). All electronic messages, images, and files are therefore by definition memes, although only a relative small minority of all such memes become socially significant. Part of the challenge of analyzing social media is revealing the socially significant memes as figure against the background of datasets. Social movements, engineered and spontaneous group dynamics (e.g., flash mobs, riots) and popular opinions are driven by the replication of messages. Such memes, therefore, can play two significant roles as currency in the marketplace of ideas: as causes of social influence processes, and as echoes of these social influence processes. First, memes in social media are causes of social influence (DeBruyn & Lilien, 2008; Harvey et al., 2011; Toole, et al., 2012; Liu-Thompkins, 2012). The second significant role that memes play is as traceable echoes that are reflective of social processes and dynamics (Chiu et al., 2007; Ratkiewicz, et al., 2011; Simmons et al., 2011). Evidence is accumulating that careful mining of tweets can reflect the extent to which flu is spreading (Nagel et al., 2013), social movement ideology is diffusing (Tsou et al., 2013a; Stefanidis et al., 2013), and how urban mobility patterns reveal geospatial and social functions (Li et al., 2013).

• Task 2A: Develop San Diego OES social media outreach and monitoring platform to distribute and analyze the spread of social media messages related to disaster alerts, emergency responses, and public announcements via different social media channels (for scenario #1).
• Task 2B: Develop computational linguistic methods (named entity extractor, hand-classified corpus, social topic-specific ontology, classifier, etc.) and social network analysis algorithms for analyzing social media diffusion patterns related to selected referendum topics (for scenario #2).
• Task 2C: Develop and implement behavioral theory analysis methods and predictive models for studying social media messages and identify the applicability of behavioral theories for understanding meme diffusion (for both scenarios).

• Task 4A: Develop HPC solutions and accessible web tools for studying human dynamics, including social media data pre-processing and data transformation tasks, ABM, spatial clustering and correlations of memes, and visualization of dynamic diffusion patterns, in accordance with the social media data structure and database systems developed by the KDC framework.