DataHalo: A Customizable Notification Visualization System for Personalized and Longitudinal Interactions

Title of the Paper

DataHalo: A Customizable Notification Visualization System for Personalized and Longitudinal Interactions

Paper Information

• Domain: Human-Computer Interaction (HCI), specifically smartphone notification management and information visualization
• Keywords: smartphone notification management, ambient information visualization, personalization, information persistence, user study, experiment-validated application, customizable display interface, graphical information encoding, notification prioritization
• DOI: https://doi.org/10.1145/3544548.3580828

Research Background and Problem

• Identified Problems:

  • People struggle to filter and prioritize smartphone notifications from different apps, particularly distinguishing notifications of varying importance from the same app.
  • Notifications lack timeliness and persistence, such as when their importance changes over time but they are easily dismissed or lost.
  • Native smartphone notification systems offer limited granularity in management, making it difficult for users to customize notification categorization or extend the accessibility of information.

• Significance:

  • With the increasing diversity and frequency of smartphone notifications, unnecessary information exacerbates cognitive load, distractions, and potential psychological stress, affecting users' efficiency and daily decision-making.

• Research Motivation:

  • To provide a flexible, user-centered solution for efficient notification management while minimizing user disruption.
  • To address gaps in existing research on notification management, particularly in the areas of personalization and long-term interaction support.

Solution

• Method or Solution:

  • Proposed a system called DataHalo, a notification visualization platform installed on an Android launcher.
  • Allows users to customize notification categorization, define importance models, and present notifications through visual encoding in an ambient visualization manner.
  • Designed keyword filtering and categorization tools combined with animated visual markers to enable persistent information reminders.

• Innovations:

  • Introduced user-centered notification categorization rules, enabling users to create virtual categories based on textual information (e.g., sender or content).
  • Supported long-term interaction through a dynamic importance model, allowing users to manage notifications more effectively based on changing importance patterns (e.g., diminishing importance over time).
  • Presented notifications with intuitive audiovisual effects (e.g., graphical markers with varying colors, sizes, positions, and animations).

• Implementation Steps and Key Technologies:

  • Notification categorization configuration: Users can define filtering rules, virtual categories, and mapping methods.
  • Information modeling: Allows users to specify dynamic patterns of notification importance over time.
  • Visualization implementation: Renders notifications as graphical markers displayed around app icons, enabling users to quickly grasp notification status.
  • Interactive features: Users can long-press icons to access configuration interfaces for diverse customization of notifications.

Research Outcomes

• Specific Results and Experimental Findings:

  • A usability study involving 17 participants found that users could understand notification categorization rules and importance models, expressing positive feedback on the system's flexibility.
  • In a three-week deployment study with 12 participants, users created 91 app configurations based on personal needs and used them to manage and categorize notifications, significantly improving satisfaction with daily notifications.
  • Data showed that user-created categorization rules differed in granularity from system-default categories (Android notification channels), with user-defined rules better aligning with individual information needs.

• Advantages Compared to Existing Solutions:

  • High customization: Enables users to create notification rules based on content and context, rather than relying on system-preset rules.
  • Information persistence: Dynamically adjusts notification importance over time, meeting users' needs for information longevity.
  • Provides a context-aware and personalized low-disruption notification management mechanism.

• Limitations and Future Directions:

  • The system only affects notification badges and does not integrate with the notification drawer or status bar.
  • There is a learning curve, such as managing parameter tuning for the importance model.
  • Challenges remain in addressing inaccuracies caused by changes in keyword labels within text.
  • Future versions could include semi-automated rule suggestion optimization and support for cross-device displays (e.g., wearable devices) to reduce interaction overhead.