DRAFT Towards Modeling Popularity of Microblogs Haixin MA 1 , Weining QIAN 1∗ , Fan XIA 1 , Xiaofeng HE 1 , Jun XU 2 , Aoying ZHOU 1 1 Institute of Massive Computing, East China Normal University, Shanghai 200062, China 2 School of Computer Science, Georgia Institute of Technology, USA Abstract As one kind of social media, microblogs are widely used for sensing the real-world. The popularity of mi- croblogs is an important measurement for evaluation of the influencial of pieces of information. The models and model- ing techniques for popularity of microblogs are studied in this paper. A huge data set based on Sina Weibo, one of the most popular microblogging service, is used in the study. First, two different types of popularity, namely number of retweets and number of possible views are defined, while their relation- ships are discussed. Then, the temporal dynamics, including lifecycles and tipping-points, of tweets’ popularity are stud- ied. For modeling the temporal dynamics, a piece-wise sig- moid model is used. Empirical studies show the effectiveness of our modeling methods. 1 Introduction Along with the development of Web 2.0 applications, social media services have attracted many users and become their hands-on toolkits for recording life, sharing ideas, and so- cial networking. Though social media services are essentially web or mobile applications and services, they combine user- generated content and social networks together, so that in- formation can be created, transmitted, transformed, and con- sumed in the cyberspace. Microblogging is one of the most important kind of social media services. Microblogging services often support users to share their information publicly through Web or mobile applications. Pieces of information, or called tweets 1) , shared E-mail: wnqian@sei.ecnu.edu.cn (correspondance author) 1) We follow the terms used by Twitter (http://twitter.com) in are usually short. A user may further retweet, or called re- post a tweet from others. Thus, data from microblogging services are often used for sensing users’ collective behav- iors [23]. Popular microblogging services including Twit- ter 2) , Sina Weibo 3) , Tumblr 4) , and etc. Popularity is an important quantative measurement for studying the influential of microblogs and authors. Model- ing the evolution of popularity for a single tweet or a set of related tweets is an essential problem for better understand- ing the collective behaviors in social media. Furthermore, the model can be used for predicating popularity of microblogs. Thus, the model can be used in applications such as online advertisement and recommendation. We study the problem of popularity modeling for a pop- ular microblogging service in China, namely Sina Weibo, in this paper. Mathematical models that capture distributions and dynamics of popularity for single and group of pieces of information are constructed. Furthermore, algorithms for es- timating parameters of those models are provided. Our con- tributions are summarized as follows. • Popularity of microblogs are formalized in this paper. We argue that the popularity should be measured by two different measurements, namely number of retweets, de- noated as #retweet, and number of possible view, de- noted as # pv. The former one is used to measure the users’ involvement for spreading information, while the latter is used to measure how many users may read the information. We show that though these two measure- ments are related, they are different aspects of the popu- larity in social media. presentation. 2) http://twitter.com/ 3) http://weibo.com/ 4) http://tumblr.com/