Quantitative Data Visualisation on Virtual Globes

Title of the Paper

Quantitative Data Visualization on Virtual Globes

Paper Information

• Subject Area: Geographic Visualization, Data Visualization
• Keywords: Geographic Visualization, Virtual Globes, Quantitative Data Visualization, User Study, Graphical Perception, Visualization Methods, Geoburst Visualization Scheme, Data Interaction Design, Human-Centered Computing, Spatial Data Visualization

Research Background and Issues

• Problems and Challenges:

  • Displaying quantitative data (e.g., global pandemic data, earthquake distribution, or population distribution) on virtual globes is a common practice, but its effectiveness has not been thoroughly studied.
  • The expressive effects, user comprehension, and comparison efficiency of different visualization methods (e.g., using 2D circles or 3D bars) remain unclear.
  • Data visualization on virtual globes faces issues such as distortion, occlusion, and interaction limitations.

• Significance:

  • With the increasing popularity of virtual globes and related interaction technologies, providing design guidelines and optimizing data visualization methods is crucial.

• Research Motivation and Related Work:

  • Existing studies mainly focus on 2D maps or localized 3D visualizations, with limited research on global 3D representations.
  • The authors aim to evaluate different visualization methods on virtual globes to explore users' perceptual differences and behaviors in real-world usage scenarios.

Proposed Solution

• Methodology:

  • Propose five data visualization methods based on virtual globes (Tangential Bars, Billboarded Bars, Normal Bars, Tangential Circles, Billboarded Circles).
  • Design a user study with three variables: angular distance between original values, relative size differences, and the maximum size of graphical elements.
  • Develop Geoburst: a novel composite visualization method combining virtual globes with radial bar charts to address issues of distortion, occlusion, and invisible data points.

• Innovations:

  • The first systematic evaluation of the comparative performance of different graphical elements (2D, 3D) and visual alignments (tangential, normal, billboarded).
  • Introduction of Geoburst, which integrates globe surface displays with radial bar charts, providing a shared scale to mitigate data distortion and occlusion.

• Implementation Steps and Key Technologies:

  • Conduct experiments using user questionnaires combined with web-based prototypes to test data reading accuracy, response time, subjective workload, and user preferences.
  • Implement Geoburst using the DBSCAN clustering algorithm to optimize bar chart arrangement and provide various user interaction options (e.g., auto-rotation, highlight linking).

Research Findings

• Specific Outcomes:

  • User Study Results: Tangential Circles had the highest error rate and perceptual workload, while 3D Normal Bars were visually appealing with minimal accuracy loss.
  • Geoburst Prototype Design: Offers methods to avoid distortion and occlusion, resulting in visualization elements that users found intuitive and aesthetically pleasing.

• Advantages:

  • Better supports accurate comparisons between data points compared to existing solutions.
  • Integrates multiple coordinate reference systems, enabling users to view hidden geographic information points.
  • Provides a method combining interactive design with quantitative representation, facilitating both global and local analysis tasks.

• Experimental or Evaluation Results:

  • Billboarded Bars and Normal Bars had the lowest error rates, with slightly longer response times but higher subjective ratings.
  • Tangential Circles had the highest subjective workload and lower accuracy.
  • Expert users gave positive feedback on the Geoburst scheme, particularly for identifying extreme values and hidden geographic distributions.

• Limitations and Future Directions:

  • Limitations:
    • Only tested relatively simple data comparison tasks; more complex pattern recognition tasks remain unverified.
    • The Geoburst scheme requires further optimization to alleviate concerns about the disruption of users' mental models caused by dynamic rearrangement of bar charts.
  • Future Directions:
    • Explore more advanced interaction technologies, such as the potential of immersive displays (AR/VR) for maps.
    • Test other projection techniques (e.g., Gilbert projection) to optimize the integration of data display and readability on globes.
    • Conduct scalability tests for varying data densities.