Scholarlink

<p>This study analyzes public sentiment surrounding the anticipated 2023 recession by examining 20,000 tweets collected over a 60-day period. Using a set of established natural language processing and machine learning techniques, we classified tweets containing recession-related keywords as positive, negative, or neutral. The results show that negative sentiment dominated the online discussion, reflecting widespread concerns about job security, inflation, and economic instability. Rather than treating social media data as predictive of macroeconomic outcomes, we frame this analysis as an exploratory case study of how economic anxieties are articulated in digital discourse. The findings illustrate both the potential and the limits of sentiment analysis for studying collective mood during periods of uncertainty. While the approach demonstrates how online communities engage with recession narratives, it also underscores challenges of representativeness, temporal scope, and reliance on older NLP models. We conclude that future work should expand to multi-platform data, adopt advanced transformer-based models, and integrate sentiment analysis with other economic and survey indicators. This reframing highlights the methodological value of sentiment analysis in capturing public discourse around crises, providing a foundation for more robust comparative studies.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Public sentiment and economic uncertainty: an exploratory analysis of Twitter discourse on the anticipated 2023 recession

  • 期刊论文

  • 出版模式：

    开放获取

  • 发表日期：

    2026年3月4日

查看全文
  • Utkansh Adlakha,
  • Sparsh Chawla,
  • Shahab Saquib Sohail,
  • Md. Tabrez Nafis,
  • Dag Øivind Madsen,
  • Gunjan Ansari

摘要

This study analyzes public sentiment surrounding the anticipated 2023 recession by examining 20,000 tweets collected over a 60-day period. Using a set of established natural language processing and machine learning techniques, we classified tweets containing recession-related keywords as positive, negative, or neutral. The results show that negative sentiment dominated the online discussion, reflecting widespread concerns about job security, inflation, and economic instability. Rather than treating social media data as predictive of macroeconomic outcomes, we frame this analysis as an exploratory case study of how economic anxieties are articulated in digital discourse. The findings illustrate both the potential and the limits of sentiment analysis for studying collective mood during periods of uncertainty. While the approach demonstrates how online communities engage with recession narratives, it also underscores challenges of representativeness, temporal scope, and reliance on older NLP models. We conclude that future work should expand to multi-platform data, adopt advanced transformer-based models, and integrate sentiment analysis with other economic and survey indicators. This reframing highlights the methodological value of sentiment analysis in capturing public discourse around crises, providing a foundation for more robust comparative studies.