BiSet: Semantic Edge Bundling with Biclusters for Sensemaking

Tomas Musil 1167504

Overview

BiSet is a visual analytics technique to support interactive exploration of coordinated relationships. Identifying coordinated relationships is an important task in data analytics. It is very hard to find coordinated relationships with existing techniques. They display individual relationships like lists of entities and many repetitious manual selections together with mental aggregations are required to find coordinated relationships. BiSet makes this whole process much easier. Coordinated relationships are modeled as biclusters. The biclusters are visualized as bundled edges between sets of related entities. These bundles enable analysts to infer task-oriented semantic insights about potentially coordinated activities.

Paper: Sun, Maoyuan, et al. "Biset: Semantic edge bundling with biclusters for sensemaking." IEEE transactions on visualization and computer graphics 22.1 (2016): 310-319.



BiSet

Features

Implementation

For the development was used JavaFX:

After loading a csv document the entities are extracted and visualized. Depending on the filters used in the gui missing components like lines and biclusters visualised.

Because of the use of Java it works on every operation system by starting the BiSet.jar file. No other installation needed.

Interaction

Entities and bicluster can be selected - their color will change. After their selection incoming and outgoing lines and other entities/bicluster they are connected to will be selected too. As often an entity/bicluster is selected as darker his color gets.

At the top the GUI is positioned. There you can change the sorting of the entities, filter bicluster regarding their # of connections or visualize either only biclusters, lines or both. On the right of the GUI you can zoom in/out, load in new csv file or clear the selection.

Links

Code documentation
Data: Tennis ATP Matches
Programm
Quellcode