Human interaction patterns on the Web over online social networks vary with the context of communication items (e.g., politics, economics, disasters, celebrities, and etc.), which leads to form unlimited time-evolving curves of information adoption as diffusion proceeds. Online communications often continue to navigate through heterogeneous social systems consisting of a wide range of online media such as social networking sites, blogs, and mainstream news. This makes it very challenging to uncover the underlying causal mechanisms of such macroscopic diffusion. In this respect, we review both top-down and bottom-up approaches to understand the underlying dynamics of an individual item's popularity growth across multiple meta-populations in a complementary way. For a case study, we use a dataset consisting of time-series adopters for over 60 news topics through different online communication channels on the Web. In order to find disparate patterns of macroscopic information propagation, we first generate and cluster the diffusion curves for each target meta-population and then estimate them with two different and complementary approaches in terms of the strength and directionality of influences across the meta-populations. In terms of the strength of influence, we find that synchronous global diffusion is not possible without very strong intra-influence on each population. In terms of the directionality of influence between populations, such concurrent propagation is likely brought by transitive relations among heterogeneous populations. When it comes to social context, controversial news topics in politics and human culture (e.g., political protests, multiculturalism failure) tend to trigger more synchronous than asynchronous diffusion patterns across different social media on the Web. We expect that this study can help to understand dynamics of macroscopic diffusion across complex systems in diverse application domains.
Keywords: bottom-up approach; macroscopic diffusion; meta-populations; synchronous diffusion; top-down approach.