Reusing Interactive Analysis Workflows

Interactive visual analysis has many advantages, but an important disadvantage is that processes and workflows cannot be easily stored reused. This in contrast to code-based workflows, which can simply run on updated datasets, adapted when necessary. In this paper, we introduce methods capture interactive visualization systems for different interactions such as selections, filters, categorizing/grouping, labeling, aggregation. These then applied making sessions reusable. We demonstrate specification using system tracks interaction provenance, allows generating from the recorded actions. The used compare versions of datasets apply them. Finally, a Python library load it directly computational notebook, providing seamless bridge between tools.