Visual Analytic

Many domains have begun collecting and processing enormous amounts of complex data in order to improve their very different processes. In many cases, event data is collected and put together as sequences. For most complex cases, multivariate events may have up to hundreds of attributes, and sequences can contain thousands of events, making the visualization of such data a real challenge. Previous research contributions have shown that such visualizations are possible, but it still remains tedious, even for experts, to go through those datasets, especially when not knowing what to look for. Based on a new classification and adaptation of existing visualization approaches and guidance strategies, we propose in this thesis a solution for how to improve the analysis of multivariate event sequences with guidance and visualization recommendations. In particular, this will be done by presenting VizREvent, a newly created application focusing on the domain of sports event analysis. Lastly, through an evaluation with both Visual Analysis experts and users, we show that our approach assists experts' and users' reasoning and question-answering capabilities and also opens up opportunities for further research.