Identifying and characterizing social media communities: a socio-semantic network approach to altmetrics

Information Science and Library Science socio-semantic network. Nodes are color-coded to identify the topics that have greater incidence. Edges are established on the basis of co-tweeted keywords. These have been filtered to a minimum of 12, and the corresponding communities are represented by overlapping areas


Altmetric indicators allow exploring and profiling individuals who discuss and share scientific literature in social media. But it is still a challenge to identify and characterize communities based on the research topics in which they are interested as social and geographic proximity also influence interactions. This paper proposes a new method which profiles social media users based on their interest on research topics using altmetric data. Social media users are clustered based on the topics related to the research publications they share in social media. This allows removing linkages which respond to social or personal proximity and identifying disconnected users who may have similar research interests. We test this method for users tweeting publications from the fields of Information Science & Library Science, and Microbiology. We conclude by discussing the potential application of this method and how it can assist information professionals, policy managers and academics to understand and identify the main actors discussing research literature in social media.

Scientometrics, 126
Wenceslao Arroyo-Machado
Wenceslao Arroyo-Machado
PhD student

My research interests include data science, altmetrics and social network analysis.