Back to top

Introduction

  • Introduction
  • Exploratory data analysis
  • Data Visualization
  • Methodology
  • Sentiment Analysis
  • Conclusion
  • Code

    • Introduction


    This study aims to systematically investigate the dissemination mechanisms and socio-psychological impacts of online rumors through computational social science methods, focusing on COVID-19-related misinformation as a case study.

    The research seeks to uncover how misinformation cascades differ functionally and affectively from factual corrections by analyzing structural propagation patterns, linguistic strategies, and sentiment dynamics.

    The ultimate goal is to generate actionable insights for three critical applications:

    • Early detection of high-risk misinformation cascades through network topology signatures;
    • Optimization of fact-checking interventions via emotion-aware rebuttals that neutralize anxiety-inducing narratives;
    • Improve platform content moderation systems by integrating propagation pattern recognition algorithms to prioritize multi-hub diffusion clusters.

    Acknowledgement of Dataset Source:
    This dataset was constructed using the Weibo API, encompassing data collected from November 2019 to March 2022. It includes 4,184 source posts and a total of 961,975 microblogs with corresponding retweets and comments, fully recording the dissemination cascades of each source post along with the authors’ user profiles. True news items were verified using official accounts, rumor-refuting posts provided by the Weibo Community Management Center, and the China Internet Joint Rumor Debunking Platform, whereas misinformation was identified via the Weibo Community Management Center’s reporting service. The dataset comprises 1,407 true news items and 768 pieces of misinformation related to COVID-19, with microblog content encoded using a TF-IDF vectorization of the top 5,000 words.