Viral Visualizations: How Coronavirus Skeptics Use Orthodox Data Practices to Promote Unorthodox Science Online

Document Title

Viral Visualizations: How Coronavirus Skeptics Use Orthodox Data Practices to Promote Unorthodox Science Online

Document Information

• Subject Area: Data Visualization, Science Communication, Social Media Studies
• Keywords: Digital Anthropology, Network Analysis, Twitter, Facebook, Data Literacy, Data Visualization

Research Background and Problem

• Problems or Challenges Identified by the Authors:
  • During the COVID-19 pandemic, social media became a battleground for debates about scientific data and policies. Mask opponents utilized orthodox scientific methods to create so-called "counter-visualizations" that promote unconventional and novel arguments.
  • Data visualization, traditionally a tool for scientific communication, has also become a site of political and social contention.
• Why This Problem Is Important:
  • Data visualization is widely regarded as an essential tool for decision support and is generally well-received by the public. However, such tools can also be used to disseminate unconventional views, undermining the credibility of mainstream science.
  • Understanding how mask opponents use data visualization to challenge public health measures can shed light on the complex dynamics of information flow during the pandemic.
• Research Motivation and Related Work:
  • Existing research often focuses on improving data visualization tools to enhance public understanding and combat misinformation. This study aims to explore how data literacy and sociopolitical contexts jointly shape public perceptions of science and policy.

Solution

• Proposed Solution by the Authors:
  • Conduct quantitative analysis of nearly 500,000 tweets related to COVID-19 data visualizations, using feature embedding and clustering methods to analyze visualization types and user networks.
  • Perform qualitative analysis of anti-mask groups on Facebook, employing digital anthropology to investigate discussion dynamics.
• Innovative Aspects of the Solution:
  • Combining quantitative techniques (e.g., network analysis, image classification) with qualitative methods from digital anthropology to deeply analyze how dissenting groups use data-driven tools to form counterarguments.
  • Introducing the term "counter-visualizations," referring to visualizations generated through orthodox scientific methods to support unconventional arguments.
• Implementation Steps and Key Techniques:
  1. Quantitative Analysis of Twitter Data:
     • Use clustering methods (e.g., k-means, UMAP) to classify visualization types based on image embedding features.
     • Construct social network graphs and detect user communities using the Louvain method.
  2. Qualitative Analysis of Facebook Groups:
     • Observe and document online activities, data discussions, and chart creation processes within anti-mask groups.
     • Apply grounded theory to label post themes and interaction patterns.

Research Findings

• Specific Findings:
  • Twitter analysis revealed that anti-mask groups frequently use bar charts, line graphs, and other "orthodox" visualization methods to argue their points.
  • Facebook research showed that while anti-mask proponents challenge mainstream scientific authority, they employ independent research and rigorous data practices to advocate for reopening the economy and opposing mask mandates.
  • The design of counter-visualizations closely resembles mainstream scientific visualizations but serves dissenting communities.
• Advantages Over Existing Solutions:
  • Moving beyond traditional theories of "lack of data literacy," this study demonstrates that these groups exhibit exceptional data analysis and visualization skills.
• Experimental or Evaluation Results:
  • Anti-mask groups not only share a large volume of visualizations but also exhibit high cohesion, fostering internal discussions and promoting their viewpoints.
  • Visualizations serve not just as tools for understanding data but also as political instruments to challenge mainstream science and policies.
• Limitations and Future Directions:
  • Limitations: The highly dynamic participation strategies of anti-mask groups may mean the study does not capture their latest tactics.
  • Future Directions: Investigate how to communicate the inherent uncertainties in science and incorporate social and cultural contexts into data visualization design.

Conclusion and Implications

1. Anti-mask proponents redefine the role of data visualization in scientific debates through advanced data literacy practices, positioning it as a core tool for political movements.
2. Data literacy is not a standardized skill but varies significantly across different social contexts.
3. Scientists and visualization researchers must recognize that visualizations are not merely unidirectional information tools but are critical arenas for sociopolitical interaction. Future research should focus on the social, historical, and multidimensional impacts of data communication with the public.