On the Identification of Influential Topics in the Social Sciences Using Citation Analysis

Abstract

In this study, we investigate the influential topics of publications in the field of Social Sciences using a novel citation analysis technique. These publications are indexed by Scopus, spanning the period from 2014 to 2022, and our analysis utilizes a graph-partitioning based topic modelling approach utilizing the citations of a publication, specifically the publication's cited papers titles. We focus on the top 1 % publications per year, based on citation count, considering the citation count metric as a concrete indicator of the influence a publication has on the community. While the study's findings reveal intriguing insights into the impact of the pandemic and the steady increase of innovation management, our primary contribution lies in the methodology we propose and utilize. This methodology can be readily adapted to examine other fields of study and can accommodate diverse types of publication information, such as abstracts, keywords, and more. Our approach involves constructing a bipartite graph that links publication IDs with their corresponding cited publications' titles keywords. By projecting this graph onto the keyword space, we identify and visually present the underlying topics that emerge from the data. This visualization offers a clear representation of the identified topics, enhancing the overall understanding of the research landscape in the Social Sciences field.