Spatio-temporal Dynamics and Network Diffusion of Weibo Sentiment Toward U.S. Tariff Policies: A Geographic Perspective
This study analyzes 1.3 million Weibo posts to reveal how U.S. tariff policies in early 2025 triggered a structured, multi-peak sentiment evolution where geographic economic factors explained up to 57.2% of local sentiment variance and distinct user groups drove the breadth and depth of information diffusion.
Original paper licensed under CC BY 4.0 (https://creativecommons.org/licenses/by/4.0/). This is an AI-generated explanation of the paper below. It is not written or endorsed by the authors. For technical accuracy, refer to the original paper. Read full disclaimer
When a government announces a major change in trade rules, such as a new tax on goods from another country, the reaction is rarely just a line on a stock chart. It is a wave of human feeling that ripples through society, changing in intensity and direction depending on where people live and who they talk to. For decades, economists have tracked these shifts using slow-moving data like monthly trade reports, which often arrive long after the public has already formed an opinion. Today, however, a different kind of sensor exists: the social media feed. Platforms like Weibo in China act as a real-time nervous system for the nation, capturing millions of individual reactions the moment news breaks. By combining the study of how people feel with the study of how information travels, a researcher can now map the geography of public emotion, seeing not just that people are upset, but exactly where the anger is hottest, how it spreads from person to person, and whether the source of the news changes how it is received.
In early 2025, a new wave of trade tensions emerged when the United States imposed a universal tax on Chinese goods, a move that affected hundreds of billions of dollars in exports. To understand how this economic shock translated into public sentiment, a researcher turned to the vast digital landscape of Weibo. They gathered a massive collection of 1.3 million posts related to the tariffs, a task that required sifting through a noisy ocean of text to find the meaningful signals. Because the posts were written in a mix of formal policy language and casual internet slang, the researcher could not rely on simple word counts. Instead, they used a sophisticated computer program, trained by a powerful artificial intelligence, to read and understand the emotional tone of each message. This allowed them to classify millions of posts with high precision, distinguishing between fear, anger, and calm, and to attach a geographic location to nearly 230,000 of them.
The researcher then watched how the conversation evolved over time, breaking the six-month period into three distinct chapters based on the rhythm of the news. They found that public attention did not simply rise and fall in a single smooth curve. Instead, it surged in sharp peaks whenever a new policy was announced, followed by a gradual settling as the initial shock wore off. The first phase, when the tariffs were first introduced, saw a rapid spike in negative feelings as uncertainty took hold. As the conflict escalated in the second phase, the volume of discussion reached its highest point, but the conversation broadened to include debates about counter-measures and economic relief. By the third phase, as tensions began to ease, the emotional temperature of the country cooled, returning toward a baseline. This pattern showed that the public does not react to a single event, but rather to a sequence of unfolding developments, with the emotional response shifting as the story changes.
What made this study particularly revealing was its look at where these feelings came from. The researcher discovered that the intensity of public sentiment was not the same everywhere; it was deeply tied to local economic realities. In provinces where the economy relied heavily on exporting goods to the United States, the reaction was far more intense and negative than in regions less exposed to trade. By using a statistical method that accounts for these local differences, the researcher found that economic factors could explain more than half of the variation in how people felt across the country. This means that a person's reaction to a national policy was not just a matter of personal opinion, but was significantly shaped by their local job market and the health of their regional industries. The data showed that the map of public emotion was a direct reflection of the map of economic vulnerability.
The study also peeled back the layers of how this information traveled through the network, revealing a clear division of labor between different types of users. When official state media accounts, which are verified with a blue checkmark, shared news, the information spread incredibly wide but stayed shallow, reaching many people in a single step without much further discussion. These accounts acted as broad broadcasters, setting the agenda for the entire country. In contrast, when individual influencers or regular users shared a post, the information traveled differently. It moved deeper, passing through multiple layers of friends and followers, creating long chains of conversation. The researcher found that negative emotions were the fuel for these deep chains; angry or worried posts were much more likely to be passed along through several degrees of separation than positive or neutral ones. This suggests that while official voices can tell everyone what is happening, it is the grassroots network of ordinary people and local influencers that determines how deeply those feelings take root and how far they travel through the social fabric.
Ultimately, the research paints a picture of a society where digital public opinion is co-constructed by two powerful forces: the hard economic realities of local regions and the hierarchical structure of social networks. The findings show that while the government and official media can broadcast a message to the entire nation instantly, the true depth of the public's emotional response is driven by local economic pain and the organic, multi-layered sharing habits of everyday users. The study confirms that in the modern era, understanding a national crisis requires looking at both the map of the economy and the map of the network, as the two are inextricably linked in shaping how a nation feels about its future.
Drowning in papers in your field?
Get daily digests of the most novel papers matching your research keywords — with technical summaries, in your language.