Modelling cross--‐network information diffusion: enhancing the understanding of social networks

One of the challenges for policy makers is not just the design of policies but how to ensure that the general public accepts these new policies. One of the barriers to acceptance is the spread of information – whether accurate, inaccurate or blatantly false – via social networks, particularly social media. (For example, vaccination programs can be significantly disrupted by anti--‐vaccination campaigners.)

The aim of this project is to develop a framework for modelling cross--‐network information diffusion. This will be used to explore the impact of social networks – both virtual and “real life” – on policy implementation, looking particularly at the case of policies surrounding infection control. This builds upon previous work on agent--‐based modelling of epidemiology.

There are two key questions that must be resolved in this investigation:

1) What is the impact of super--‐imposed networks (e.g. local community, work community, Facebook friends, Twitter, etc.) on information diffusion? Current studies in network--‐based information diffusion focus on single social networks, such as Twitter or Facebook, without recognising that there is cross—network information flow that could influence the diffusion. This aspect of the project will draw upon the sociological literature on information diffusion, using this to create a computational framework for cross—network interactions. The plausibility of this model will be evaluated by extending examples of information diffusion that have been studied a single network.

2) What are the best ways of providing feedback on social network information diffusion to support policy makers? It is one thing to develop a model of this information diffusion, but in order for it to be useful to policy makers, it must be understandable. Detailed overviews of the social network, which are typically provided by agent--‐based social simulation models, may not be appropriate. Alternatives, such as a more game--‐like mechanism (e.g. a virtual social media analyst) could provide a more meaningful experience. This aspect of the project would involve development and evaluation of different presentations of the model data to users (the policy makers and analysts).