DigitalTraces: Unveiling Fraud through Interactive User Behaviour Exploration
Loading...
Date
2025
Journal Title
Journal ISSN
Volume Title
Publisher
The Eurographics Association
Abstract
Fraud detection teams in financial institutions face the challenge of identifying suspicious activity within user behaviour. However, existing tools often lack the ability to seamlessly integrate multiple dimensions of digital activity into a single, interactive visualisation, leading to increased cognitive load and preventing analysts from quickly spotting anomalies in varying sources of information. This paper introduces DigitalTraces, a visual analytics tool aimed at improving the detection of fraudulent patterns particularly in the dimensions tied with digital activity. The system combines several stacked timelines to offer an overview of multiple activity dimensions, integrating online banking session data, device identifiers, transactional activities, and account information. We validated our tool with a think-aloud experiment where two fraud analysts were tasked with detecting anomalies in a financial fraud scenario. Experts emphasised the tool's ability to provide intuitive insights and enhance understanding.
Description
CCS Concepts: Human-centered computing → visualisation systems and tools
@inproceedings{10.2312:evs.20251085,
booktitle = {EuroVis 2025 - Short Papers},
editor = {El-Assady, Mennatallah and Ottley, Alvitta and Tominski, Christian},
title = {{DigitalTraces: Unveiling Fraud through Interactive User Behaviour Exploration}},
author = {Narciso, João Bernardo and Feliciano, Beatriz and Costa, Rita and Bizarro, Pedro},
year = {2025},
publisher = {The Eurographics Association},
ISBN = {978-3-03868-282-0},
DOI = {10.2312/evs.20251085}
}