The Digital Bibliography and Library Project (DBLP) is a popular computer science bibliography website hosted at the University of Trier in Germany. It currently contains 2,722,212 computer science publications with additional information about the authors and conferences, journals, or books in which these are published. Although the database covers the majority of papers published in this field of research, it is still hard to browse the vast amount of textual data manually to find insights and correlations in it, in particular time-varying ones. This is also problematic if someone is merely interested in all papers of a specific topic and possible correlated scientific words which may hint at related papers. To close this gap, we propose an interactive tool which consists of two separate components, namely data analysis and data visualization. We show the benefits of our tool and explain how it might be used in a scenario where someone is confronted with the task of writing a state-of-the art report on a specific topic. We illustrate how data analysis, data visualization, and the human user supported by interaction features can work together to find insights which makes typical literature search tasks faster.
The DBLP contains more than 2 million paper titles. More advanced visualization techniques are required which provide additional views on keywords, authors, and also the paper content. Nobody is able to get an overview about this vast amount of textual data - neither the experts nor the novice researchers like students being new in a field of research.