RaDIA: Visual Analytics for Systematic Sample Association in Nuclear Weapon Test Verification Workflows
Our late-breaking report "RaDIA: Visual Analytics for Systematic Sample Association in Nuclear Weapon Test Verification Workflows" (PDF) has been accepted at EICS 2026 in Patras, Greece. This is work by Stian Verherstraeten together with Christophe Gueibe (SCK CEN), Gustavo Rovelo Ruiz, and myself.
National Data Centres (NDCs) that monitor the Comprehensive Nuclear-Test-Ban Treaty face a recurring analytical challenge: determining whether anomalous radionuclide detections at different stations and times share a common source. Atmospheric transport can spread radioactive particles over thousands of kilometres, so what looks like several independent signals may in fact trace back to a single release event. Today, analysts compare source-receptor sensitivity (SRS) fields one by one across separate CTBTO tools and document suspected links in personal spreadsheets.
RaDIA (Radionuclide Data Integration and Analysis) brings these views into a single coordinated workspace. A spatial overlap detection algorithm formalizes what analysts do manually: it counts shared grid cells in backward atmospheric trajectories above a dilution threshold, producing quantitative association scores. Results are shown as an interactive map, a temporal Sankey diagram where link width encodes overlap strength, and a sortable table. Progressive disclosure guides analysts through the workflow step by step, keeping cognitive load manageable while preserving full control over parameters.
Citation
@inproceedings{verherstraeten2026radia,
author = {Verherstraeten, Stian and Gueibe, Christophe and {Rovelo Ruiz}, Gustavo and Luyten, Kris},
title = {{RaDIA}: Visual Analytics for Systematic Sample Association in Nuclear Weapon Test Verification Workflows},
booktitle = {Companion Proceedings of the 18th ACM SIGCHI Symposium on Engineering Interactive Computing Systems},
series = {EICS Companion '26},
year = {2026},
location = {Patras, Greece},
publisher = {ACM},
address = {New York, NY, USA},
doi = {10.1145/3807968.3810929}
}