Title: Bridging Spatial and Social Network Analysis through Visual Analytics

Date: Monday, May 11, 2026

Time: 4:30–6:30 PM Eastern time (U.S.)

Location: Technology Square Research Building (TSRB) Room 334 (VIS Lab) 

Virtual Meeting (hybrid): https://gatech.zoom.us/j/6902674214?pwd=WXgrZ1RkOVlhaWhUci83R1h0UER1QT09

Sichen Jin

Ph.D. Candidate in Computer Science

School of Interactive Computing, College of Computing

Georgia Institute of Technology 

Committee

Dr. Clio Andris (Advisor) - School of Interactive Computing, School of City & Regional Planning, Georgia Institute of Technology

Dr. Alex Endert - School of Interactive Computing, Georgia Institute of Technology

Dr. John Stasko - School of Interactive Computing, Georgia Institute of Technology

Dr. Yalong Yang - School of Interactive Computing, Georgia Institute of Technology

Dr. Shrirang Abhyankar - Principal - AI & Optimization, The Electric Reliability Council of Texas

Abstract
Spatial social networks (SSNs) represent social relationships embedded in geographic space, where nodes and edges are associated with geographic locations. Spatial social network analysis (SSNA) requires methods from both social network analysis (SNA) and geographic information systems (GIS), with applications in domains such as criminal network investigation, geosocial science, epidemiology, transportation and mobility analysis, urban planning, and location-based services. However, existing tools and workflows remain fragmented: SNA often ignores the influence of underlying geographic context, while GIS tools lack support for modeling relational structures. As a result, analysts often rely on disconnected pipelines that hinder effective SSNA.

This dissertation addresses these challenges by investigating how visual analytics can bridge SNA and GIS for integrated SSNA through design studies, long-term empirical evaluation, and pedagogical implementation. These research goals and contributions are organized into four main thrusts:

(1) A visual analytics system for integrated SSNA. We present SNoMaN, an interactive visual analytic system that supports simultaneous exploration of network and geospatial contexts through coordinated views, interactions, and dynamically computed spatial–network metrics. The system enables users to examine how spatial distributions relate to network structures, such as identifying whether densely connected communities are geographically clustered or dispersed. It also provides interactive visualizations for exploring relationships between geographic metrics (e.g., Euclidean distance, spatial dispersion, and average connection distance) and network metrics (e.g., degree, betweenness and closeness centralities, network density, and shortest path length), as well as for identifying anomalies.

(2) Uncovering interdisciplinary needs and challenges. We conducted a workshop-based study using SNoMaN as a design probe with researchers from SNA and GIS backgrounds. The findings reveal both the value of visual analytics as a boundary object for cross-disciplinary collaboration and key sources of science friction, including differences in terminology, analytical expectations, and familiarity with spatial versus network concepts. These insights informed iterative refinement of the system.

(3) Empirical evaluations of integrated SSNA. Building on iterative refinements of the system, we evaluated SNoMaN through longitudinal case studies with 12 domain experts using real-world datasets across diverse application domains. Over multiple sessions spanning several months, the results demonstrate that SNoMaN effectively supports exploratory spatial data analysis, hypothesis generation, and insight discovery. At the same time, the study highlights limitations related to transparency and trust in computed metrics, suggesting the need for improved interpretability and interoperability with programmable tools.

(4) SSNA teaching modules using visual analytics. We developed a learner-centered SSNA teaching module built around SNoMaN, designed to lower barriers to interdisciplinary learning across SNA and GIS. The curriculum was implemented across eight universities in multiple disciplines and evaluated through instructor observations, student engagement, and learning outcomes, leading to insights for broader interdisciplinary education.

Overall, this work shows that integrating geographic and network perspectives through visual analytics enables new forms of reasoning about SSNs that are difficult to achieve with fragmented workflows. By combining design study, empirical evaluation, and pedagogical contributions, this dissertation contributed to the use of visual analytics for interdisciplinary SNA and GIS analysis and supports broader adoption of SSNA in both research and education.