Observing the evolution of social network on Weibo by sampled data

Abstract

Although there have been many researches on the online social networks (OSNs), observing the evolution of a real OSN is still interesting and instructive for understanding people's behavior in OSNs. In this paper, the actual evolution of the social graph of a real OSN - Weibo, is studied by sampled data. The exact timestamp of creating or removing each following relationship cannot be sampled. However, by the created time of the users' accounts, the evolution of the social network of Weibo is roughly observed. In this way, it is found that the growing pattern of the network scale shows S-shape. Some other properties of the network, such as network density, the number of connected components, the efficiency of the network, clustering coefficient, degree assortativity, and so on, are also observed. As the network grows, the density of the network keeps reducing and eventually reaches a steady state. The change of the number of connected components indicates the users' crowd behavior during the network evolution.