Visualization overview: Using modern text mining techniques to provide insight into visualization research practice

We take a new approach to analyze what the general focus of Visualization research is, among the panoply of different topics and approaches. Due to the intrinsic characteristics of this kind of in-depth research, being often too time-consuming and difficult to carry, we resort to text mining techniques to streamline this analysis. This study was carried out by applying topic modeling for discovering the abstract topics that occur in visualization papers. The study used the vispubdata.org dataset as the reference to gather almost every paper presented, from 1990 to 2018, at the IEEE Visualization (VIS) set of conferences: InfoVis, SciVis, VAST, and Vis. We requested ten topics and assigned each one its ten most important and representative terms. With this analysis, we intend to envelop the current practices in the visualization research community.