Visualization, analysis and structural pattern infusion of DBLP co-authorship network using Gephi

Abstract

DBLP (Digital Bibliography & Library Project) has the huge collection (around 3.4 million) of journal articles with its meta data, papers published in various national and international conferences, and other number of online publications in the field of computer science. Research data on DBLP increases enormously which arises many research problems in the domain of bibliography data network analysis such as citation index and co-authorship etc. For this purpose, many researchers' work is available on Data visualization, Data Mining, Data Analysis and prediction of the same. Here in this research paper, we are showing results of visualizing and analyzing DBLP Co-authorship network with the help of Gephi tool. We use DBLP dataset which is an undirected graph and contains 17280 nodes and 58539 edges. In this Graph nodes are taken as author and edges connect two nodes/authors according to their association in considered research papers' dataset. There will be more than one edge in between two nodes if both the researcher co-authored more than one paper. Attention has been devoted towards statistic based analysis with some varying measuring factors such as connected nodes, clustering coefficient, page rank, Average path length and Graph density. We apply Force Atlas2 for community generation and thereafter graph represent co-authorship network communities. We discuss and detail about collect, analyze and visualize co-authorship data which is further used for understanding the academic collaboration and finding the research communities for specific knowledge domain. Even we discuss about analysis of individual researcher and their community in the past through repetitive co-published work.