ParcoursVis
Progressive Visualization of Patient Pathways at Scale
ParcoursVis is a visual analytics system designed to help experts explore patient care pathways at scale. Developed in collaboration with the Aviz Team and the French Social Security, it supports the analysis of how treatments are actually delivered across France and how they compare with official recommendations.
The project addresses a major challenge in healthcare data analysis: millions of patients and potentially billions of events must be processed and visualized efficiently. To make this possible, ParcoursVis relies on progressive data analysis, aggregating large datasets incrementally so that users can interact smoothly with the visualization without noticeable delays.
By transforming raw healthcare event data into interactive visual representations, ParcoursVis enables experts to investigate treatment pathways at national or regional scale, opening up forms of analysis that were previously difficult to perform.
The project is funded by the Health Data Hub and AP-HP, and was originally initiated through a collaboration with CMAP / École Polytechnique and CNAMTS.
Live demo | Project page | Source code
People Involved
- Ambre Assor
- Jean-Daniel Fekete
- Sungbok Shin
- Mickaël Sereno
- Qing Chen
- Catherine Plaisant
- Emmanuel Bacry
- Stéphane Gaiffas
Publications
-
Ambre Assor, Jean-Daniel Fekete
Understanding Expert Exploration in EHR Visualization Tools: The ParcoursVis Use Case
2025 IEEE 16th Workshop on Visual Analytics in Healthcare (VAHC), pp. 1–7
DOI: 10.1109/VAHC69430.2025.00005 -
Ambre Assor, Hyeon Jeon, Sungbok Shin, Jean-Daniel Fekete
Changing the Paradigm from Dynamic Queries to LLM-generated SQL Queries with Human Intervention
arXiv preprint arXiv:2509.09461, 2025
Project Images