RidgeBuilder: Interactive Authoring of Expressive Ridgeline Plots

Research Background and Issues

• Identified Problems or Challenges:
  • Ridgeline plots are an important visualization method for displaying the distribution or evolution characteristics of multiple datasets. However, most tools currently only support creating basic Ridgeline plots, lacking support for diverse layouts and styles.
  • Adjusting the layout of Ridgeline plots (e.g., overlap degree and order) significantly impacts the representation effect, but fine-tuning these layouts and styles often requires complex design software like Adobe Illustrator, which imposes high skill requirements on users.
  • Existing code-driven tools and interactive tools provide plotting functionality but are limited. Code-based tools (e.g., ggridges and Vega-Lite) are flexible but have a steep learning curve, while interactive tools (e.g., Tableau) are user-friendly but lack options for complex layouts and decorations.
• Importance:
  • Ridgeline plots are widely used in both scientific and artistic fields, effectively expressing time series evolution, multivariate distributions, and more, showcasing strong data visualization value.
  • Creating more flexible and expressive Ridgeline plots can expand their application scenarios and help users extract more valuable insights.
• Research Motivation and Related Work:
  • Current tools lack sufficient exploration of the design space for Ridgeline plots and fail to efficiently address layout optimization issues.
  • This study aims to liberate Ridgeline plots from the constraints of traditional tools, enhancing their expressive capabilities and inspiring more creative applications.

Solution

• Proposed Method or Solution:
  • Introduce an intuitive design-space-based tool—RidgeBuilder—for interactively creating expressive Ridgeline plots.
  • Design a five-dimensional design space for Ridgeline plots: elements (e.g., axes, plot types), layout (e.g., overlap, hierarchy), style (e.g., color, transparency), decoration (e.g., annotations, background), and composition.
  • Solve layout adjustment issues through an optimization algorithm for Ridgeline layouts, proposing three design goals:
    1. Adjacent curves should exhibit similar peak patterns.
    2. The curvature of Ridgeline curves should be minimized.
    3. Layouts should avoid obscuring major peaks.
• Innovative Aspects of the Solution:
  • Propose an optimization model that formalizes the Ridgeline layout problem as a Traveling Salesman Problem (TSP) to achieve real-time layout optimization.
  • Develop a specific three-axis data model to enhance the expressive flexibility of Ridgeline plots.
  • Provide style template recommendations, enabling users to easily experiment with classic styles (e.g., "vinyl record style").
• Implementation Steps and Key Techniques:
  • Combine user interaction with core functionalities (e.g., data binding, layout adjustment, style modification, decoration addition) to make the Ridgeline plot generation process more intuitive.
  • Use algorithms to automatically generate high-quality layouts and recommend optimal layout choices.
  • Interface design supports intuitive actions like drag-and-drop and slider adjustments, while allowing export to SVG files for advanced editing.

Research Outcomes

• Specific Outcomes:
  • RidgeBuilder can create diverse and expressive Ridgeline plots, significantly reducing user operation complexity compared to existing tools.
  • Provides an optimization algorithm for interactive layout adjustments, enabling users to more effectively showcase data evolution patterns.
• Advantages Over Existing Solutions:
  • Compared to code-driven tools (e.g., ggridges), RidgeBuilder is more user-friendly for non-expert users, shortening the learning curve.
  • Compared to traditional interactive tools (e.g., Tableau), RidgeBuilder offers greater flexibility and innovation in layout selection and style control.
• Experimental or Evaluation Results:
  1. User Reproduction Study: Test participants scored RidgeBuilder with an average SUS score of 88.75/100, unanimously indicating ease of use and efficiency.
  2. Algorithm Evaluation: The optimized layout algorithm significantly outperformed baseline methods (e.g., statistical sorting), achieving the highest user evaluation score (77.33), confirming the effectiveness of the three proposed design goals.
  3. Time Efficiency: The algorithm achieves real-time responsiveness (<1 second) and high-quality experience (most scenarios completed within 10 seconds).
• Limitations and Future Directions:
  1. Limitations:
     • Does not yet support combining Ridgeline plots with other chart types.
     • Algorithm response speed may require optimization for very large datasets.
     • Evaluation could not fully compare other tools fairly, as most lack specific functionalities for Ridgeline plots.
  2. Future Directions:
     • Introduce human-machine collaborative design with dynamic optimization suggestion modules.
     • Enhance direct manipulation support for charts (e.g., freehand trendline drawing).
     • Integrate with existing visualization tools (e.g., as a Figma plugin) to expand usage scenarios.

Through RidgeBuilder, the design and implementation of Ridgeline plots become more intuitive and powerful, paving the way for extensive applications in research, art, and business. This study may inspire more researchers to explore the potential and expressiveness of specific visualization forms like Ridgeline plots.