ZipLine: New visual tool bridges graph topology and attributes for deeper analysis
A predicate language lets analysts query graphs by both structure and data at once.
Deep Dive
ZipLine is a visual analytics system for multivariate graphs that bridges topology and attribute spaces using a unified predicate language. A predicate-learning algorithm translates user interactions into formal expressions, enabling iterative reasoning across both spaces. Demonstrated on energy infrastructure, cybersecurity, and drug discovery, ZipLine supports expressive pattern discovery that existing systems struggle with.
Key Points
- Introduces a predicate language that unifies topology and attribute queries in multivariate graphs
- Includes a learning algorithm that translates user interactions into formal predicates across both spaces
- Demonstrated via three case studies: energy infrastructure, cybersecurity, and drug discovery
Why It Matters
Makes complex graph pattern discovery practical for cybersecurity, energy, and pharma analysts without sacrificing depth.