An Image-based Global Climate Sentiment Index

Area of Science:

• Environmental science
• Climate change research
• Social science

Background:

• Climate change impacts necessitate understanding public sentiment for policy and risk management.
• Existing methods relying on text data face language dependence and cross-country comparability issues.

Purpose of the Study:

• To develop a globally comparable measure of climate sentiment.
• To overcome limitations of text-based sentiment analysis by incorporating visual data.
• To provide a robust tool for monitoring climate attitudes and market expectations.

Main Methods:

• Collected ~3.19 million climate-related images from 19 countries (2015-2024) via Google Images.
• Detected positive and negative sentiment within visual data.
• Constructed a multi-frequency Climate Sentiment Index.

Main Results:

• Developed a novel Climate Sentiment Index using visual data.
• Achieved greater accuracy and cross-cultural robustness in sentiment identification.
• Created a long-horizon, globally comparable dataset.

Conclusions:

• The visual-data-driven Climate Sentiment Index offers a powerful, language-independent tool.
• This index enhances understanding of public climate attitudes and market expectations.
• Applicable across economic, social, and environmental domains for informed decision-making.