Visualization and pattern discovery of social interactions and repost propagation in Sina Weibo

Many researchers focus on analysis of social interactions and information propagation, but they do not provide results in an intuitive way that helps people understand easily. In this paper, we aim to visualize social interactions of a group of users and repost propagation in Sina Weibo, and seek for patterns from visualization. We propose a new definition to formulate the relationship between users, which reflects their social interactions. We then visualize the data with the help of multidimensional scaling. The result shows an interesting pattern that social interactions among people change corresponding to significant events. Moreover, we visualize repost propagation of hot weibo posts in Sina Weibo. We discover that repost propagation is explosive in the very first time period. And authorized user accounts are helpful for faster and wider propagation. We later choose another popular repost case to support our finding.