Analyzing Collective Intelligence Through Sentiment Networks in Self-organized Douban Communities

Understanding the communication behaviors of online community users, particularly their crowd intelligence dynamics, has long been a focal point within network communication research. In this study, we present an approach integrating the BERTopic topic model, advanced Natural Language Processing (NLP) techniques, and Social Network Analysis, to meticulously dissect the intricate dynamics of emotion propagation and evolution within collective behaviors that unfold on social networks. Our investigation delves deeply into this complex landscape, exploring the relationships between the sentiment of the initial post and subsequent responses, the interplay between sentiment strength and activity levels, and the correlation between sentiment polarity and the intensity of activity. This study highlights the significance of harnessing the combined power of BERTopic, NLP, and social network methodologies to decode the subtleties of emotional propagation and transformation.