Nakazawa, RinaItoh, TakayukiSaito, TakafumiJohansson, Jimmy and Sadlo, Filip and Marai, G. Elisabeta2019-06-022019-06-022019978-3-03868-090-1https://doi.org/10.2312/evs.20191179https://diglib.eg.org:443/handle/10.2312/evs20191179We usually use a text-based search engine while surveying research papers. Such search systems have difficulties for novice researchers in case they do not know appropriate keywords or do not understand the positions of papers. Many visualization tools of citation networks have been proposed to help this task. These tools demonstrated that not only text information of papers but citation relationships and co-author relationships also are helpful clues for research survey. We propose CoCoa, a linked network visualization of co-citation and co-author relationships for surveying research papers. Our system visualizes both citation and co-author networks at the same time. To make comparison and grasp of correspondence between co-citation and co-author networks easier, the system treats both a paper and an author as bags of words and cluster them into topics applying LDA (Latent Dirichlet Allocation) at the same time. Based on the clustering result, it places the clusters of a citation network by a hybrid force-directed and space-filling algorithm. The position of topic clusters in the networks would have an influence on the correspondence of a particular topic in the networks. Our system extracts the clusters which consist of the common combinations of topics in two networks. Then it reuses the positions of the clusters in a citation network as the initial cluster positions of a co-author network, supposing there are a large number of authors.Humancentered computingVisual analyticsVisualizationInformation visualization10.2312/evs.20191179109-113