Structural summaries for visual provenance analysis

Many systems today exist to collect provenance that describes how some data was derived. Such provenance represents useful information in many use cases, e.g., reproducibility or derivation process quality. Depending on the use case, collected provenance traces need to be explored and analyzed. Therefore, various approaches, including visual analysis approaches have been proposed. However, these typically focus on analyzing individual provenance traces. We propose to create structure-based summaries of provenance by aggregating many provenance traces provided in W3C-PROV representation. We further describe the analysis tasks that apply on these summaries. We showcase the usefulness of structural summaries based on several use cases, when using appropriate visualization and interaction techniques.