Repository relative to our paper ZigzagNetVis: Suggesting temporal resolutions for graph visualization using zigzag persistence
Raphaël Tinarrage, Jean R. Ponciano, Claudio D. G. Linhares, Agma J. M. Traina, and Jorge Poco.
IEEE Transactions on Visualization and Computer Graphics 31, 1–18 (2025)
In data analysis, temporal graphs constitute a powerful framework for modeling dynamic and complex systems. In this context, our work has focused on two questions:
- the automatic suggestion of temporal resolution (or timeslice length), a crucial parameter at the basis of many analysis techniques,
- the visual representation of the temporal graph, allowing the user to explore the data and detect patterns or anomalies.
We adress both problems from the perspective of Topological Data Analysis, and more specifically zigzag persistent homology.
▶ claudiodgl.github.io/ZigzagNetVis
A backend-free build of the ZigzagNetVis frontend that runs entirely on GitHub Pages — no server. The colored zigzag-persistence barcodes and the merge/split flows are precomputed into static JSON; the node-link diagrams are sliced and laid out live in the browser. Double-click the barcode to drop the three timestamp markers (t(2) black, t(1)/t(3) at t±10) and open three node-link diagrams; drag the markers to inspect them, switch resolution, toggle the merge/split flows, and isolate categories or connected components with coordinated highlighting across all views.
Source for the static demo: github.com/claudiodgl/ZigzagNetVis (bundled networks: Primary School, Hospital, High School).
Link to the full prototype (server-based): http://visualdslab.com/zigzagnetvis/
A Jupyter notebook, including our implementations of resolutions suggestion and colored barcodes, is found at https://github.com/raphaeltinarrage/ZigzagNetVis/blob/main/Demo.ipynb
An example of the ZigzagNetVis system using the Primary School network.
