Exploiting coordinated views for scholarly reading and analysis

There are several ways of reading scientific papers, and in general, documents. The sequential reading of the text flow can be complemented with alternative visualizations such as tree-based views, maps and charts. Infoview techniques are very promising in this context but not yet fully exploited. This paper presents a system called DocuDipity that is able to smoothly integrate infoview techniques, in particular a SunBurst visualization and a flow text reader, in a coordinated environment that helps researchers to read and analyze scientific articles. We discuss some end-user tests that showed that our tool is highly effective to find patterns and disclose some habits in writing articles. For instance, we used DocuDipity to study how the authors organise their content and use citations. The paper presents both the system and the details of the evaluation, together with some indications on future developments.