Comparing Community-based Information Adoption and Diffusion Across Different Microblogging Sites

The proliferation of social media is bringing about significant changes in how people make sense of their world and adopt new information. However, social, cultural and political divisions continue to separate users and information into different social media systems. Twitter and Facebook, for example, are strictly forbidden in mainland China. As a result, 21.97% of all world-wide Internet users are thus excluded from participating in these platforms. In this study, we investigate whether the dynamics of information diffusion, modeled as the adoption patterns of topical hashtags, differ between the communities of the mentioned social media sites as a result of this separation. Specifically, we compare Weibo and Twitter, the two largest micro-blogging sites serving respectively the Chinese population and the rest of the world, by exploring the similarities and differences of how their respective users adopt new information. By leveraging sophisticated community detection algorithms and heterogeneous graph mining methods, we investigate and compare how the different characteristics of these communities influence information diffusion and adoption. Experimental results show that while community-specific information influences topic diffusion and adoption in both environments, novel features, extracted from heterogeneous graph based communities, have a greater effect on Weibo information adoption than Twitter. We also find that users sharing hashtags is an important factor in information diffusion on both Twitter and Weibo, whereas user mentions are important for Weibo, but less so for Twitter. Overall, we conclude that Weibo and Twitter differ sharply in how their users adopt information in response to similar factors.