SGV: Spatial Graph Visualization
Published in ACM SIGSPATIAL 2025 — Minneapolis, MN, 2025
Authors: Tarlan Bahadori, Alvin Chiu, Ahmed Eldawy, Michael Goodrich
SGV is a scalable spatial graph visualization framework that retains the geographic meaning of networks while producing clean, uncluttered layouts. It integrates classical force-directed forces with novel spatial anchoring—a gravity-based pull toward real-world regions or points—yielding layouts that respect both topology and geography.
Why SGV
Traditional force-directed layouts disregard spatial context, while purely geographic maps ignore graph structure. SGV introduces a unified anchored force model that preserves both, reducing clutter and improving readability in large, dense networks. The system is implemented on Apache Spark, enabling layouts at massive scale.
Key Capabilities
- Anchored force model combining attraction, repulsion, and spatial gravity
- Region- and point-based anchors for flexible geographic constraints
- Scalable Spark implementation for graphs with millions of vertices
- Improved readability (shorter anchor distances, fewer crossings, lower HEL)
Artifacts:
- 📄 Paper: /files/sgv_sigspatial25.pdf
- 🎥 Demo Video: Watch on Google Drive
- 💻 Code: github.com/tarlaun/FDGV