Interactive Analysis of the Discussion from a Virtual Community on Neuroscience

Abstract

Social media on the Internet has been promoting disruptive transformations in the society, enabling new possibilities for knowledge construction. Nowadays, researchers can fastly gather a large number of discussions from a virtual community of interest for analysis. This paper presents the study of a relevant community on neuroscience, with more than 43,000 registered members. The research method employs a process based on Grounded Theory and Knowledge Discovery in Databases (KDD), using a tool crafted to support interactively and iteratively use of data mining algorithms such as topic modeling and sentiment analysis. From the analysis of 2,927 posts and 19,227 comments, the results reveal the most prominent subject regards Alzheimer’s disease, followed by general acknowledgments and requests for help whether concerning symptom assessment, examination results analysis, and medical advice. Most of the identified topics have a positive polarity, indicating that interactions are predominantly friendly. Negative feelings emerged from controversial topics, being mostly non-technical or speculative subjects such as mind control techniques, help with MRI results, answers to medical advice, and theories about consciousness. The findings reinforce the feasibility of such studies and show useful insights regarding the community interest in neuroscience.