INTUITIVE: Information Theory for User-Guided Exploration
In Visual Analytics, the analysis process can consist of multiple views generated by versatile interactions with the system. The multiple views are necessary since we have to provide the user with the possibility to analyze the data via interaction in different ways to gain insights. The interaction should be goal-oriented, i.e. it should lead to insights to satisfy the information need of a user. This is the case if only the relevant data is shown to the user to help him to confirm or reject a hypothesis.
During the interaction with the system, a user does not get feedback and has to rely on her intuition, if she is on right path to answer a question.
Which is the correct path that a user has to take and can we identify if this path is optimal or are we able to identify redundant steps?
Think about ways to measure the information content when interacting with various datasets. What methods are used nowadays to measure information?
Create meaningful visualizations to communicate the change of information content along the analysis process. Exemplary questions are:
How can we visually support the user in finding out whether an interaction with the system is expedient (goal-oriented) or not?
Can we give evidence whether certain interactions with a system were redundant?
Is it possible to suggest/recommend alternative interactions, which might be more useful with respect to a specific task?
Design a prototypical framework to get intuitions about these questions.