D3: A Massive Dataset of Scholarly Metadata for Analyzing the State of Computer Science Research

Proceedings of the 13th Conference on Language Resources and Evaluation (LREC 2022), pages 2642–2651 Marseille, 20-25 June 2022 © European Language Resources Association (ELRA), licensed under CC-BY-NC-4.0 2642 D3: A Massive Dataset of Scholarly Metadata for Analyzing the State of Computer Science Research Jan Philip Wahle†, Terry Ruas†, Saif M. Mohammad††, Bela Gipp† †University of Wuppertal Germany, ††National Research Council Canada †{wahle, ruas, gipp}@uni-wuppertal.de ††saif.mohammad@nrc-cnrc.gc.ca Abstract DBLP is the largest open-access repository of scientific articles on computer science and provides metadata associated with publications, authors, and venues. We retrieved more than 6 million publications from DBLP and extracted pertinent metadata (e.g., abstracts, author affiliations, citations) from the publication texts to create the DBLP Discovery Dataset (D3). D3 can be used to identify trends in research activity, productivity, focus, bias, accessibility, and impact of computer science research. We present an initial analysis focused on the volume of computer science research (e.g., number of papers, authors, research activity), trends in topics of interest, and citation patterns. Our findings show that computer science is a growing research field (≈15% annually), with an active and collaborative researcher community. While papers in recent years present more bibliographical entries in comparison to previous decades, the average number of citations has been declining. Investigating papers’ abstracts reveals that recent topic trends are clearly reflected in D3. Finally, we list further applications of D3 and pose supplemental research questions. The D3 dataset, our findings, and source code are publicly available for research purposes. Keywords: Computer Science, Scientometrics, Research Trends, NLP, DBLP, AI 1. Introduction In the last few decades, computer science (CS) has transformed many scientific fields. Faster systems, more accurate results, and efficient tools are just some of the benefits provided by computer science advance- ments. Arguably, today there is hardly any area not affected by its vast possibilities. Consider how diffi- cult it would be to test, develop, and research new vac- cines without access to tools of informatics (e.g., public repositories). The techniques behind these contributions are often made available through scientific publications which can be used to investigate trends and patterns within computer science itself. However, computer science is a large field with many sub-areas (e.g., natural lan- guage processing (NLP), computer vision) and reposi- tories (e.g., arXiv); thus, a complete analysis of its pub- lications is a challenging task. How fast is computer science research growing? How many authors are ac- tively publishing in their field? What topics are preva- lent in specific venues? A large and carefully curated dataset of CS-publications metadata is crucial for the quantitative exploration of these questions. DBLP is one of the largest open computer science repositories with records from major journals and pro- ceedings starting from 1936.1 The repository provides access to several pieces of metadata associated with each of its papers, including author names, title, year, and venue. The papers stored in DBLP come from paid publishers (e.g., IEEE, ACM) and open reposi- tories (e.g., ACL, arXiv). NLP Scholar (Mohammad, 2020c) and arXiv also offer an …

Content truncated to 500 words