Leading Journals in ChatGPT Articles with More References Co-citations Using Slope Graphs in R with Web R-Platform: Bibliometric Analysis (Preprint)

Abstract

BACKGROUND ChatGPT has sparked numerous innovative applications in healthcare, with PubMed currently indexing 2,692 articles on the topic. However, there is no available information on which journals have more references co-citations in ChatGPT articles and their thematic features. OBJECTIVE This study aims to answer three key questions: (1) How can major themes be extracted from ChatGPT-related articles to respective journals? (2) Does the number of Chinese publications in ChatGPT merely lag behind that of American authors? (3) Which journals have more references co-citations in ChatGPT articles, and what characteristics do these articles share based on ChatGPT keywords? METHODS We collected 2,489 articles with "ChatGPT" in their titles from the Web of Science core collection, gathering metadata from 2022 onwards. To analyze these articles, we employed various visualization methods, focusing on slope graphs to compare trends in prestigious journals based on three aspects: the overall number of publications, the top 100 cited articles, and references co-citations on journals. We created performance sheets, network charts of author collaborations and co-words for 2,489 articles, and an impact beam plot for the top 100 cited articles. Additionally, we used the follower-leading clustering algorithm (FLCA) for clustering analysis, which efficiently identified research focal points and emerging trends. RESULTS The study revealed the following findings: (1) Three major themes emerged from the top 20 keywords: ARTIFICIAL-INTELLIGENCE, IMPACT, and QUALITY; (2) The number of ChatGPT research publications by Chinese authors still merely lagged behind those by US authors; (3) The prestigious journals leading in references co-citations among ChatGPT articles were Nature and the Journal of Medical Internet Research(JMIR); (4) Both journals were classified as ARTIFICIAL-INTELLIGENCE in nature. CONCLUSIONS The study identified three major themes (ARTIFICIAL-INTELLIGENCE, IMPACT, QUALITY), confirmed US dominance in ChatGPT research over China, and highlighted Nature and JMIR as leading journals in references co-citations among ChatGPT articles, both focusing on artificial intelligence. Based on the results of this study, slope graphs created in R on web R-platform can be applied to research, not just ChatGPT articles.