Analysis of Scientific Publications from Turkey in Dermatology: Social Network Analysis, Text Mining, and Clustering

Abstract

Objective: To perform network analysis of publications from Turkey under the Dermatology category of the Web of Science (WoS) between 1998 and 2017. Material and Methods: We selected all SCI-E journals listed in the WoS under the "Dermatology" category. For descriptive analyses, Clarivate Analytics of WoS was used. VOSviewer was used for text mining, co-word clustering, and visualizing the obtained results. Results: A total of 3319 articles from Turkey were published over the span of 20 years. There is an increasing trend of publications under the "Dermatology" category until 2006, after which point the number seems to stabilize. Between 2002 and 2017, the average percentage of articles from Turkey is around 3.5% in the Dermatology category of WoS. Five large medical institutions such as Istanbul University, Gülhane Military Medical Academy, Hacettepe University, Ankara University, and Ankara Numune Hospital produced 30% of the dermatology articles from Turkey. The average number of citations per article is 9.5. The number of citations shows an increasing trend, reaching almost 3,500 in 2017. We obtained five clusters by co-authorship analysis. Generally, we have seen that geographic distance is an important factor for collaboration of institutions. Via text mining, 43,030 terms were obtained. After processing, the most important 64 terms were collected under five clusters. Conclusion: The dermatology research output of Turkey is increasing, although it has slowed in the last decade. An important fraction of these publications originate from older institutions in Istanbul and Ankara. The publications from Turkey are both qualitatively and quantitatively concentrated on psoriasis and Behçet's disease.