We conducted an investigational study of the formulation of the heterarchical online knowledge-based community among university students, which also involved users outside a course. As an exercise in a course, students were assigned to post their opinions regarding global issues on Twitter to connect with social actors. The emerging all connections were collected by Twitter's application programming interface. Dataset was categorised into types of behaviourgraphics, which were styles of online users' individual behaviour, as proposed by Solis. There were 954 tweets by 197 users, and 13 within 20 types of behaviourgraphics were exhibited in the Twitter community. However, students' isolated tweets without any connection occupied 74% of all tweets. The tweets of social actors counted for merely 14% of students' tweets, and connections in the community proved to be sparse. Compared with the types of behaviourgraphics in the results of our previous study, we could identify the following problems in students' messages: content without relational cues, inhibition points in the content, lack of communicative behaviour, insufficient consideration of timeline, and lack of likeability. The results indicate the need to cultivate students' sociability skills and their pre-investigation of other communities in order to expand their knowledge-based community. In terms of future studies, we discuss the importance of promoting a bridge to enhance connections with another community.