Next-Generation Interactive and Immersive Graphs: A Unity-Based Framework for Advanced Data Visualization

Abstract

Over the past decades, data visualization has significantly evolved from traditional two-dimensional (2D) representation of text-based data towards immersive three-dimensional (3D) visualizations, leveraging emerging technologies such as Virtual Reality (VR). Within this advancement, social network graphs that represent entities as nodes and relationships as edges, essentially enabling connection analysis, have attracted particular interest for analyzing and understanding complex connections. Although research focusing on these graphs has concluded to numerous tools and applications, a framework that incorporates multiple smart techniques to improve user experience and be broadly applicable to different age and users’ background groups, while utilizing an efficient and interactive approach of visually separating relationship types remains relatively limited. Various frameworks aim to enhance immersion and usability through selected user experience techniques, however, there remains a lack of an unified solution that combines a wide range of features including custom data import, multi-device support, diverse and intuitive interaction methods, and an interactive filtering mechanism designed to fit within the visualization environment. This master’s thesis proposes a Unity-based framework designed for immersive 3D and intuitive visualization of social network graphs, with broad applicability to desktop and VR devices, ensuring an enjoyable and engaging user experience. The framework integrates functionalities and intelligent techniques inspired by previous studies that aim to enhance information clarity and provide a user-friendly experience, distinctively employing the Fibonacci lattice layout algorithm, to evenly distribute nodes around in an imaginary sphere and reduce visual clutter in crowded networks. In addition, we propose an immersive filtering approach that incorporates color-coded 3D elements to represent different relationship types, allowing users to interactively reveal or hide specific connections within the network. This method offers a more intuitive and visually engaging alternative to distinguish different relationship types instead of traditional checkbox-based filtering panels. To assess the framework’s effectiveness, a user study is conducted, where participants’ accuracy and task completion time is assessed through specific use cases, followed by a questionnaire to gather insights into their overall experience, usability, and perception of this filtering technique.