Agnostic Visual Recommendation Systems: Open Challenges and Future Directions.

Area of Science:

• Computer Science
• Human-Computer Interaction
• Data Visualization

Background:

• Visualization Recommendation Systems (VRSs) aid users in generating insightful data visualizations.
• Existing systems often rely on human-defined rules, limiting autonomous learning.
• Agnostic VRSs aim to imitate human analysts for data relationship identification and chart design.

Purpose of the Study:

• To summarize the current literature on agnostic VRSs.
• To identify key challenges hindering the progress of agnostic VRSs.
• To outline future research directions for agnostic VRSs.

Main Methods:

• Literature review of agnostic VRSs.
• Analysis of common obstacles in agnostic VRS development.
• Identification of research gaps and future opportunities.

Main Results:

• Agnostic VRSs show high application potential but face significant hurdles.
• Key obstacles include lack of standardized datasets, difficulty in learning design rules, and defining perceptual effectiveness criteria.
• Progress in autonomous chart generation is currently limited by these factors.

Conclusions:

• Agnostic VRSs represent a promising but challenging research area.
• Addressing the identified obstacles is crucial for advancing the field.
• Future research should focus on developing standardized datasets, robust learning mechanisms for design rules, and objective evaluation metrics.