"Not just alluvial" for Better Visual Comparison Among Set Partitions
At the conference "Vision, Modelling and Visualization (VMV) 2024", we presented our paper titled "Not Just Alluvial: Towards a More Comprehensive Visual Analysis of Data Partition Sequences", authored by Madhav Poddar, Jan-Tobias Sohns, and Fabian Beck. It extends the stream-based alluvial diagrams for better visual comparison among partitions of sets.
This paper contributes to Project CoCo by enabling the analysis of clustering result sequences. However, it is domain-agnostic, making it applicable to broader set of use cases, such as evaluating multi-label classifier models.
Paper:
https://diglib.eg.org/items/0c36bb49-db0f-4d45-96c4-3745edacc86b
GitHub Repository:
https://github.com/madhavpoddar/not_just_alluvial