SetVis: Visualizing Large Numbers of Sets and Intersections

Set-type data occurs in many domains such as life sciences (Lamy & Tsopra, 2019), health (Landolfi et al., 2022) and the retail industry (Adnan & Ruddle, 2018), as well as in generic applications such as analysing structures of missing data (Ruddle, Adnan, & Hall, 2022) and association rule mining (Wang et al., 2020). SetVis is new matrix-based set visualization software, implemented as a Python package which is available from PyPi. The documentation is available from setvis.readthedocs.io which contains various hands-on Jupyter notebooks guiding users to use the package. SetVis uses a memory-efficient design, operates with datasets held in RAM or a database such as PostgreSQL, and allows user interaction graphically or programmatically. A technical evaluation shows that SetVis uses orders of magnitude less memory than the UpSet (Lex et al., 2014) Python package (Nothman, 2022).