Influence analysis for celebrities via microblog platform

The exponential propagation of online social networks has re-shaped the way that how people acquire information, post messages, and interact with each other. A growing population of social users, especially those celebrities and entertainment stars, tend to maintain their social relationships and build their reputation over such a type of emerging platforms. In this paper, we aim to analyze the social influence of entertainment stars by tracking their social information. Taking Sina Weibo, the most popular microblogging platform in China, as a trial platform, we propose an unified framework to crawl microblogs associated with celebrities and analyze these messages to assess their social influence. In particular, this framework consists of three procedures: 1) We propose a cross-media method to acquire celebrities' social accounts. 2) We implement a distributed architecture to crawl microblogs periodically, while meeting the access limits. 3) Based on the acquired information, we undertake various types of analysis in terms of fans count, fans geographical distribution, microblog contents, microblog posting frequency, microblog propagation, and social relationship graph. Currently, we have crawled a set of microblogs related to more than 10,000 celebrities, and built a real system at NTU campus to present insights to potential stakeholders, such as celebrities, ad agencies.