Garima Jindal supervised by Prof. Kamalakar Karlapalem received her Doctorate in  Computer Science and Engineering (CSE). Here’s a summary of her research work on Scalable Visual Analytic Solutions for Node and Community Insights in Large Networks

Visualizing large networks is essential for understanding complex systems. However, designing scalable network visualization methods remains a significant challenge. Traditional node-link representations often suffer from severe visual clutter due to overlapping nodes and edges, whereas matrix-based representations face issues related to space usage and readability. Although several techniques have been proposed, such as hybrid representations (e.g., NodeTrix, ChordLink) and details-on-demand exploration, to enhance the visual analysis of large networks, the problem remains largely unsolved. A key analytical task in such networks is to identify central or influential nodes within and between communities, which have applications ranging from marketing to epidemiology. Although centrality measures provide quantitative estimates of node importance, they often lack context on the influence of a node within its own community and in the broader network. We identify a critical gap in the literature: the absence of intuitive and scalable visualizations that enable users to explore both node centrality and community properties in large networks. To address this, the thesis introduces a suite of novel visual analytic techniques designed to support intuitive and scalable network exploration. We proposed and implemented a Spiral Visualization that facilitates users to visually comprehend large networks. Spiral visualization is a representation that highlights key aspects of networks, including the number, size, and density of communities, important or central nodes within communities, centrality distribution within communities, connections between communities, and connections between nodes. To facilitate analysis and comprehension of networks using various interaction techniques, such as zooming, tooltip, and highlight, we have implemented a Spiral Visualization dashboard. We conducted a qualitative user study that incorporated observation, think-aloud protocols, and the participant’s confidence and ease to assess the usability and suitability of our visualization. The findings suggest that our visualization is appropriate for the evaluated network tasks. However, tasks requiring color comparison, such as identifying the densest community and comparing community densities, were found to be more challenging to perform. In addition, its scalability is limited to networks of approximately 10,000 nodes. To overcome these limitations, we introduce the Scalable Spiral Visualization, an enhanced version capable of visualizing networks with up to 40,000 nodes. This improved design addresses key shortcomings of the original spiral layout, such as dense node placement near the center and overlapping spirals caused by large community sizes. Like the original, the scalable version enables visual exploration of both node-centric and community-centric properties. Usability is further enhanced through an interactive dashboard that supports zooming, filtering, tooltips, and node highlighting. The effectiveness of the technique is validated through a case study and an ICE-T evaluation, both of which support the design choices. To further overcome the non-linear spatial limitations of spiral layouts, we develop a linear visualization approach. This method arranges communities along a vertical axis based on importance and places nodes within them in a linear layout. The linear visualization is designed to be user-friendly and scalable, allowing clearer community separation and easier visual interpretation of node significance. A comparative user study was conducted to assess the relative strengths and usability of the linear and spiral visualizations. The results show that while spiral visualization is more familiar due to its resemblance to traditional node-link diagrams, linear visualization offers superior usability, especially in identifying and comparing important communities. In summary, this thesis contributes scalable, interactive, and intuitive visual analytics techniques for the exploration of large networks. By allowing for an effective visual interpretation of the centrality of the node and the community properties, the proposed methods help users extract meaningful information with a reduced cognitive load. 

June 2026