WebGraphViz: A WebGL-Based Interactive Graph Visualization Tool for Visual Data Management System (VDMS)
Luis Aguilar
Department of Computer Science
Faculty Supervisor: Shahrukh Humayoun
Machine Learning (ML) techniques increasingly rely on extracting information from large volumes of visual data (images, videos, etc.). Intel Lab’s Visual Data Management System (VDMS) addresses this issue by storing multiple visual data formats in a high-performance graph database. However, extracting insights from VDMS is challenging due to the lack of interactive visual interfaces that allow the user to discover patterns. Due to this issue, this work proposes WebGraphViz, an interactive visual analytics tool that leverages WebGL for GPU-accelerated rendering performance and real-time exploration. Similarly, using retail sample data, we designed an interactive visualization to help users analyze customer journeys and reveal patterns within VDMS’s graph data. Lastly, we conducted performance evaluations across three distinct interaction experiments in both high- and low-performance computing environments. The results demonstrate the significant performance improvements achieved by WebGraphViz compared to GraDVis, another visual interface for VDMS, validating its effectiveness for the interactive exploration of large-scale graph data.