Visual Encodings of Temporal Uncertainty: A Comparative User Study.

Area of Science:

• Information Visualization
• Human-Computer Interaction
• Cognitive Psychology

Background:

• Uncertainty visualization is crucial for data interpretation.
• Representing temporal uncertainty poses unique challenges due to time's characteristics.
• Existing methods for visualizing uncertainty may not be optimal for temporal data.

Purpose of the Study:

• To compare the effectiveness of various visual encodings for temporal uncertainty.
• To identify optimal visualization techniques for different temporal uncertainty tasks.

Main Methods:

• A comprehensive study was conducted comparing six visualization types: gradient plots, violin plots, accumulated probability plots, error bars, centered error bars, and ambiguation.
• Participants performed tasks involving uncertain start and end times.
• Error rates and task completion times were measured for each visualization and task.

Main Results:

• Significant differences were observed in error rates and completion times across visualization types and tasks.
• Ambiguation and error bars were more effective for judging durations and temporal bounds.
• Gradient plots were more effective for judging probability values.

Conclusions:

• The choice of visualization technique significantly impacts the accuracy and efficiency of interpreting temporal uncertainty.
• Ambiguation or error bars are recommended for tasks focused on temporal extents.
• Gradient plots are recommended for tasks focused on probability estimation within uncertain temporal intervals.