DysVis: A User-Centred Data Visualization System for Dyslexia Pre-screening

Research Background and Issues

• Issues and Challenges

  • Dyslexia is a common neurobiological learning disorder that significantly impacts an individual's reading, writing, and spelling abilities.
  • Most current pre-screening tools focus primarily on Latin-based languages, providing insufficient support for students whose primary language is Chinese.
  • Assessing dyslexia in Chinese presents unique challenges: the logographic nature of Chinese characters, complex stroke order, and physical motion data involved in the learning process are difficult to incorporate into evaluation methods.
  • Existing pre-screening tools often lack the ability to deeply analyze students' learning processes and instead focus on delivering final screening results.

• Importance

  • Early identification of dyslexia is crucial for effectively helping students overcome learning difficulties and realizing their potential, especially in regions with limited educational resources (e.g., in Chinese contexts).
  • Data-driven support tools that employ analysis methods tailored to specific linguistic and cultural contexts can significantly improve screening accuracy.

• Research Motivation and Related Work

  • This study proposes designing a user-centered data visualization system for students in Chinese contexts to meet the specific needs of special education teachers in screening students with dyslexia.
  • By adopting multiple evaluation dimensions, including handwriting data, keypoint analysis of physical motion, and behavioral insights, this study aims to address the current gaps in pre-screening tools.

Solution

• Methods and Solutions

  • This study developed a user-centered data visualization system named DysVis, which integrates handwriting analysis, physical motion keypoint transformation, and a user-friendly visualization interface to meet the needs of special education teachers.
  • The system's core functionalities include real-time handwriting animation display, transformation and visualization of physical motion data, and multi-layer analysis interfaces (ranging from overall student performance to specific behavioral insights).

• Innovations

  • Focusing on dyslexia identification in Chinese contexts, the system incorporates unique dimensions related to Chinese learning, such as stroke order observation and behavioral feature capture.
  • For the first time, the system combines teacher feedback to design a four-tier panel: student overview, task overview, sub-question analysis, and behavioral analysis, providing targeted support for early intervention and educational guidance.

• Implementation Steps and Techniques

  1. User Needs Assessment:
     • Conducted semi-structured interviews with 13 special education teachers to refine design requirements. Key needs include cross-disciplinary student performance evaluation, rapid identification of learning difficulties, and in-depth analysis of specific learning issues.
  2. System Design:
     • Integrated the following components:
       • Handwriting Data Analysis: Presented real-time handwriting animations using SVG to reveal stroke order and spatial issues.
       • Physical Motion Keypoint Transformation: Processed video data using OpenPose technology to extract physical motion and skeletal keypoints.
       • User Interface Design: Multi-layer visualization panels to help teachers intuitively identify problems through data.
  3. System Development and Iteration:
     • The system underwent iterative optimization through multiple rounds of teacher participation, incorporating feedback to enhance data insights and interface usability.

Research Outcomes

• Specific Outcomes

  • DysVis provides multi-layer data visualization interfaces, including: an overview panel (displaying overall student performance), a task panel (using bar charts to delve into specific performance), a sub-question panel (observing detailed issues and handwriting animations), and a behavioral panel (capturing dynamic behavioral data).
  • With its real-time handwriting animation and motion analysis features, DysVis enables teachers to better understand students' handwriting thought processes and behavioral characteristics.

• Advantages Over Existing Methods

  • The system's targeted design can distinguish borderline cases (students whose performance is close to average but may have potential issues).
  • Compared to traditional methods, the system offers more precise and multidimensional insights (e.g., real-time handwriting dynamics and physical motion keypoints).

• Experimental or Evaluation Results

  • The system was validated through user research, including case studies, user studies, and teacher interviews.
  • Over 88% of participating teachers found DysVis intuitive enough to quickly identify students with potential learning difficulties and assist in developing personalized educational strategies.
  • In terms of user experience, DysVis's interface design was rated highly for ease of use, with handwriting animations and bar charts significantly aiding teachers in quickly identifying problematic students.

• Limitations and Future Directions

  • Limitations:
    • The study involved a limited sample size of students (only 4 confirmed dyslexia cases), necessitating further expansion of the sample pool to improve system applicability.
    • Data privacy poses potential challenges for video and motion capture functionalities; future research could explore alternatives like heatmap cameras with stronger privacy protection capabilities.
    • The current system is primarily limited to traditional Chinese contexts and needs validation for its applicability to other writing systems (e.g., simplified Chinese).
  • Future Directions:
    • Expand the system's application to analyze students' strengths and weaknesses, supporting broader personalized teaching.
    • Introduce machine learning algorithms to enhance the prediction capabilities for students with potential dyslexia.
    • Optimize interface design, such as providing teachers with more concise summary reports to help focus on problem areas quickly.

Through this study, DysVis offers a comprehensive and effective solution for applying educational technology in special needs education. This design approach can also be extended to other complex language issues and learning disability domains.