GLLA: A Unified Force-Directed Graph Layout Framework Supporting Local Adjustments

Area of Science:

• Computer Science
• Graph Theory
• Data Visualization

Background:

• Force-directed algorithms are standard for node-link diagrams but lack local adjustment capabilities.
• Uniform force application hinders customization for specific graph structures, like dense clusters.

Purpose of the Study:

• Introduce a unified force-directed graph layout framework supporting local adjustments (GLLA).
• Enable fine-grained control over graph layouts for enhanced visual perception.

Main Methods:

• Developed two unified parametric force expressions.
• Integrated a unified force model with an improved stochastic gradient descent solver.
• Designed GLLA to reproduce existing algorithms and enable local modifications.

Main Results:

• GLLA matches mainstream algorithms in structure preservation and readability.
• Demonstrated notable improvements in time performance compared to existing methods.
• Validated GLLA's ability to meet diverse local adjustment needs across four real-world scenarios.

Conclusions:

• GLLA offers a flexible and efficient solution for force-directed graph layout.
• The framework successfully balances global layout properties with local customization.
• GLLA enhances the practical application of graph visualization techniques.