IGScript: An Interaction Grammar for Scientific Data Presentation

Literature Title

IGScript: An Interaction Grammar for Scientific Data Presentation

Literature Information

• Subject Area: Scientific data visualization and interactive data presentation
• Keywords: Data presentation, interaction grammar, scientific visualization, dual-space view, script generator, domain-specific language, data storytelling

Research Background and Problems

• Problems or Challenges:

  • Existing scientific visualization solutions primarily focus on the computation or rendering stages, lacking support for the data presentation stage.
  • Techniques based on general-purpose programming languages (GPLs) are often difficult to master, resulting in a steep learning curve for most domain experts.
  • Creating smooth and visually appealing transition animations, such as scientific presentation videos or data storytelling, is challenging for domain experts to accomplish manually.
  • Defining complex logical queries (e.g., combinations of 3D box queries) using traditional graphical user interfaces (GUIs) is inefficient and prone to errors.

• Significance:

  • Data presentation is a crucial aspect of scientific data depiction and interpretation, supporting domain experts in exploration, data analysis, and science communication.
  • Interactive data presentation enables smooth visual transitions between multiple important data features, enhancing communication effectiveness.
  • Editable animation segments provide significant support for adjustments or new requirements in a user-friendly manner.

• Research Motivation and Related Work:

  • Most existing scientific visualization methods are strongly dependent on current programming language components, leading to syntax and code styles that are highly tied to the host language.
  • Few works provide editable and highly intuitive tools for data presentation, especially for complex domain-specific data operations.

Solution

• Methods and Solutions:

  • Propose IGScript, a script-based interaction grammar tool that supports the generation of interactive data presentation animations using natural language-style scripts.
  • Design a lightweight compiler to convert script syntax into data animations, along with a code generation tool (decompiler) to support the transformation of interactive exploration into script code.
  • Develop a dual-space linked interface that integrates visualization space and coding space, enabling real-time interaction and script customization.
  • The script supports general data loading, object transformation, animation control, and scene transitions, while also providing application-specific data processing, such as vector field data clustering and DTI fiber bundle operations.

• Innovations:

  • The development of an independent compiler makes the tool free from reliance on existing programming languages, reducing the learning difficulty for domain experts.
  • The script syntax supports quick and intuitive data presentation functionality, resembling natural language and suitable for non-programmers.
  • The dual-space view provides real-time visual feedback and fine-grained editing, combined with syntax parameter assignment modules to achieve comprehensive interactive presentation.
  • Specialized syntax design for complex scientific data needs, such as 3D box logical queries and dynamic tracking.

• Implementation Steps and Key Technologies:

  1. Dual-space linked interface: Users define coarse-grained ROIs in the visualization space and fine-tune parameters in the coding space.
  2. Compiler and script generator: Convert natural language-style scripts into animation presentations while recording interactions to generate corresponding code.
  3. Editable presentations: Support the trimming, copying, deletion, and merging of presentation scripts to output customized videos.

Research Outcomes

• Specific Outcomes:

  • User testing indicates that IGScript is user-friendly for domain experts and general audiences, particularly excelling in defining ROIs, generating animations, and editing animation segments.
  • Demonstrated the usability and flexibility of IGScript through various scientific datasets, including 2D scalar fields, 3D scalar fields, vector fields, and DTI data.
  • Performance evaluations show extremely short compilation times and high operational efficiency, making it suitable for real-time use.

• Advantages:

  • Script code is easy to understand, significantly reducing the learning and operational difficulty for domain experts.
  • The editability of data presentation animation segments greatly enhances the applicability and flexibility of traditional methods.
  • Supports advanced domain-specific data operations, optimizing the handling of complex logical combinations and interaction needs.

• Experiment or Evaluation Results:

  • Participating users gave high ratings for tool design engagement and effectiveness, especially praising the ROI definition and expression query functionalities.
  • The results generated in data storytelling animation cases clearly showcased scientific data features and supported domain discussions.

• Limitations and Future Directions:

  • Current ROI definitions rely on user interaction; plans to introduce machine learning algorithms for automatic recommendations.
  • Need to expand support for a broader range of data types and domain-specific syntax while considering enhancements to code editing experiences through GUI flowcharts.
  • Intend to extend applications to science communication, remote scientific data collaboration, and introduce more types of data interactions and device support, such as VR/AR.

In summary, IGScript successfully addresses the customization and editability needs of domain experts in scientific data presentation by combining a lightweight compiler and interactive grammar tools, providing a groundbreaking solution for scientific visualization and data storytelling creation.