From Invisible to Visible: Impacts of Metadata in Communicative Data Visualization

Area of Science:

• Human-Computer Interaction
• Information Visualization
• Cognitive Psychology

Background:

• Context is crucial for understanding data visualizations.
• Invisible context, such as missing metadata, can hinder transparency and comprehension.
• The impact of recontextualizing visualizations with metadata is not well understood.

Purpose of the Study:

• To investigate how chart type, topic, and user goals influence the perceived relevance of different metadata categories.
• To examine the effects of metadata on visualization transparency, trustworthiness, relevance, and understanding.
• To determine if metadata impacts the accuracy and recall of information from visualizations.

Main Methods:

• Experiment 1: 64 participants evaluated real-world visualizations, selecting desired metadata types based on given goals.
• Experiment 2: 144 participants analyzed visualizations with and without metadata, assessing comprehension, transparency, relevance, and trustworthiness.
• Statistical analysis of participant responses to quantify the impact of metadata.

Main Results:

• Participants prioritized encoding explanation metadata for understanding goals and source metadata for trustworthiness assessment.
• Visualizations with metadata were perceived as more thorough but not significantly different in relevance, accuracy, clarity, or completeness.
• Metadata did not affect the accuracy of extracted information but influenced the salience of remembered details.

Conclusions:

• Metadata significantly impacts user perception of visualizations, particularly regarding thoroughness and trustworthiness.
• The type of metadata most valued is goal-dependent, highlighting the need for tailored contextual information.
• While metadata doesn't improve factual accuracy, it can shape memory and perceived importance of visualization content.