Visual text analysis techniques have been one of the most prominent topics over the past decades. Many researchers have recently been interested in using text visualizations techniques and tools to conduct the structural analysis of a text. Text visualization indeed uses graphical representations to help text analysts to draw more knowledge from a raw text and to easily uncover the relations between the terms of a text whether explicitly or not. However the analyst’s interactivity on the visualizations is determinant for a deeper understanding of the text. This work aims at improving the user interaction on the visualizations provided by EVOQ, an existing tool for text structural analysis. We have implemented a new algorithm for EVOQ which not only generates visualizations based on a pre-encoded terms relation dictionary but also helps analysts to construct the visualizations at ease. After evaluating our proposed solution with some users and the existing visualization interaction techniques, we have suggested some guidelines for future improvements.
|la date de réponse||2 sept. 2019|
|Superviseur||Bruno Dumas (Promoteur) & Isabelle LINDEN (Copromoteur)|