Wed, 01. Aug. 2018   Latif, Shahid

Profile Pages for Scientists

Natural language generation meets visualization – a novel approach of presenting publication records

How can we assess the experience and expertise of a researcher? In addition to the formal resume, publication records play an important role in answering this question. The analysis of these records provides insights into researchers’ careers such as their publication statistics, evolution of research topics, coauthor relationships, and collaboration groups. However, the reporting of this extracted information in a comprehensive yet easy-to-understand manner is a challenge. We suggest an unorthodox and novel approach to use automatically generated text in combination with visualizations to convey such information in a clear and self-explanatory way. We argue that the close integration of text and visualizations not only provides explicit explanations but also supports the exploratory analysis through interactions. We have developed VIS Author Profiles (VAP), a Web-based visual analytics tool, that generates profiles of researchers in the form of interactive reports. VAP can help hiring managers to recruit research personnel, researchers to find collaborators, students to find potential supervisors, and research program committees to recruit members. The VAP system is accessible online. The corresponding paper has been accepted to be published in the renowned journal IEEE Transactions on Visualization and Computer Graphics (TVCG) and will be presented at IEEE VIS 2018 in Berlin, Germany.