LiterMiner: A Literature Visual Analytic System

The data set of scientific literature contains abundant knowledge. Currently, most bibliography search engines provide only keyword search or similarity search functions for retrieving information directly stored in the database but neglect mining hidden information. In this article, we design a visual analytic tool called LiterMiner to extract entities such as article, author, affiliation, and keyword from these literature records, establish the relationships among them and display them to help analyst understand them effectively. Especially, LiterMiner uses community detection algorithm to explore academic teams. Our tool has three principle features. First, a sophisticated data clean process is designed for integration of many different formatted literature records. Second, LiterMiner can find academic teams and analyze the resume of an academic team. Third, an interactive and cooperate visualization solution is used to display the analysis result, which help user understand the result better.