Consensus and Contradictions: A Cross-Organizational Analysis of Visualization Style Guides

Honorable Mention
Interactive Data VisualizationData StorytellingVisualization Perception & CognitionUI/UX DesignersData Scientists & AnalystsHCI Researchers

Paper Title

Consensus and Contradictions: A Cross-Organizational Analysis of Visualization Style Guides

Publication Info

  • Topic area: Systematic analysis of visualization style guides across organizations and industries.
  • Keywords: Visualization style guides, data visualization, design norms, organizational practices, consensus, contradictions, chart-specific guidelines, chart-agnostic guidelines, sociotechnical artifacts, value-sensitive design.

Background and Problem

  • Problem / challenge: Visualization style guides, while influential in shaping data communication, are inconsistent and lack systematic analysis across organizations and industries. Prior studies have been limited in scope and fail to address cross-organizational patterns or the sociotechnical values embedded in these guides.
  • Significance: Style guides influence how data is presented to the public, impacting clarity, trust, and decision-making. Understanding their consensus and contradictions can improve visualization practices and bridge gaps between academic research and professional application.
  • Motivation and related work: Previous research has analyzed individual or small sets of guidelines, focusing on academic or practitioner perspectives. However, these studies do not provide a unified, cross-organizational view or address the values and assumptions encoded in guidelines. This paper aims to fill this gap by analyzing 53 style guides from diverse sectors.

Solution

  • Proposed approach: A systematic analysis of 53 publicly accessible visualization style guides, creating a standardized corpus of 2,120 chart-specific guidelines and developing the Guidelines Explorer tool for transparency and exploration.
  • Novelty:
    1. Curated and released a standardized corpus of 53 style guides across journalism, government, nonprofit, corporate, and academic sectors.
    2. Conducted a multi-method analysis to identify consensus, contradictions, and the sociotechnical values embedded in guidelines.
    3. Introduced the Guidelines Explorer, an interactive tool for exploring and comparing guidelines across organizations and industries.
  • Procedure and key techniques:
    • Collected and standardized style guides into a JSON corpus.
    • Used a hybrid human–AI pipeline to extract action–target units and cluster guidelines.
    • Conducted quantitative and qualitative analyses to identify patterns of consensus and conflict.
    • Annotated guidelines with value tags to uncover underlying design priorities.
    • Developed the Guidelines Explorer to enable interactive exploration of the dataset.

Results

  • Concrete findings:
    • Bar charts are the most frequently addressed chart type, with strong consensus on rules like starting axes at zero and limiting color usage.
    • Line charts show moderate agreement on practices like limiting the number of lines but diverge on zero-baseline recommendations.
    • Pie and donut charts are highly contested, with some organizations prohibiting their use and others allowing them under specific conditions.
    • Chart-agnostic guidelines emphasize color usage, simplicity, accuracy, and readability, with clarity being the most frequently cited value (92%).
  • Advantage over baselines:
    • Provides the first large-scale, cross-organizational analysis of visualization style guides.
    • Identifies shared norms and contradictions, offering insights into the values shaping visualization practices.
    • The Guidelines Explorer enables deeper exploration and comparison of guidelines, bridging gaps between research and practice.
  • Experiments / evaluation:
    • Dataset includes 53 style guides from journalism, government, nonprofit, corporate, and academic sectors.
    • Analysis covers 2,120 chart-specific guidelines and 34 value tags.
    • Hybrid human–AI pipeline validated with strong agreement metrics (e.g., Cohen’s κ ≈ 0.84 for action–target extraction).
  • Limitations and future work:
    • Corpus excludes proprietary or non-English style guides, limiting industry and geographic representation.
    • Intent extraction may oversimplify nuanced guidelines.
    • The Guidelines Explorer has not undergone formal user evaluation.
    • Future work could expand the dataset, refine NLP techniques, and track how guidelines evolve over time.

Summary

This paper systematically analyzes 53 visualization style guides, uncovering shared norms, contradictions, and the sociotechnical values embedded in organizational practices. Key findings include strong consensus on bar chart rules, fragmented agreement on line charts, and contested guidance for pie and donut charts. Chart-agnostic guidelines emphasize clarity, simplicity, and accuracy, reflecting institutional priorities. The Guidelines Explorer tool complements the analysis by enabling interactive exploration of the dataset. This work bridges gaps between academic research and professional practice, providing a foundation for improving visualization guidance and fostering transparency in design norms.

Quick Actions

Share

Share this page

ios_share

https://hci.top/en/papers/chi/222042/2026

AdRecommended

Learn AI Coding at CodeNow

open_in_newOpen DOI Link
DOI: https://doi.org/10.1145/3772318.3791628
At a Glance

Paper Snapshot

fact_check
dataset
Source
CHI
calendar_month
Year
2026
emoji_events
Award
Honorable Mention
group
Authors
1 authors
sell
Subtopics
Interactive Data Visualization, Data Storytelling, Visualization Perception & Cognition
work
Professions
UI/UX Designers, Data Scientists & Analysts, HCI Researchers
article
Content Status
Full text indexed
hub
Related Papers
10 related papers