HOLA: Human-like Orthogonal Network Layout.

Area of Science:

• Computer Science
• Graph Theory
• Human-Computer Interaction

Background:

• Extensive research has yielded numerous automatic network layout algorithms over 50 years.
• Current algorithms, despite advancements, do not match human-generated layout quality.
• Existing methods include fast heuristics for large networks and multi-stage frameworks for smaller ones.

Purpose of the Study:

• To introduce a novel human-centered methodology for designing automatic network layout algorithms.
• To address the deficiency in automatic layout quality compared to human designers.
• To develop and evaluate an algorithm based on human aesthetic criteria.

Main Methods:

• Conducting user studies to identify key aesthetic criteria for network layout.
• Developing a new automatic orthogonal network layout algorithm (HOLA) informed by these criteria.
• Performing follow-up user studies to evaluate HOLA's output against existing methods and human layouts.

Main Results:

• The HOLA algorithm, developed using the human-centered methodology, produced measurably superior orthogonal layouts.
• User studies confirmed that HOLA's layouts are preferred over those from the best available orthogonal layout algorithm.
• HOLA achieved layout quality comparable to that produced by human experts.

Conclusions:

• The human-centered methodology is effective in improving automatic network layout quality.
• HOLA represents a significant advancement in automatic orthogonal network layout.
• Future work can apply this methodology to other network layout types.