CrisisNews: A Dataset Mapping Two Decades of News Articles on Online Problematic Behavior at Scale

TL;DR

CrisisNews addresses the need to understand how online problematic behaviors originate and escalate into large-scale social media crises by building a 20-year, news-derived dataset of 93,250 articles, with a 1,354-event annotated subset. The authors define social media crises operationally, develop a scalable data pipeline (keyword filtering, GPT-4o-assisted labeling, and semantic merging), and annotate multi-dimensional factors including stakeholders, behaviors, platforms, and outcomes. Key findings reveal that regular users and group dynamics often drive crises, platforms are active actors in some cases, and viral amplification (Broken Harmony) can both harm and catalyze positive change, informing governance and safety strategies. The work provides a theoretical and empirical foundation for a comparative science of social media crises, advocating cross-platform, proactive interventions to create safer, more trustworthy online environments.

Abstract

As social media adoption grows globally, online problematic behaviors increasingly escalate into large-scale crises, requiring an evolving set of mitigation strategies. While HCI research often analyzes problematic behaviors with pieces of user-generated content as the unit of analysis, less attention has been given to event-focused perspectives that track how discrete events evolve. In this paper, we examine 'social media crises': discrete patterns of problematic behaviors originating and evolving within social media that cause larger-scale harms. Using global news coverage, we present a dataset of 93,250 news articles covering social media-endemic crises from the past 20 years. We analyze a representative subset to classify stakeholder roles, behavior types, and outcomes, uncovering patterns that inform more nuanced classification of social media crises beyond content-based descriptions. By adopting a wider perspective, this research seeks to inform the design of safer platforms, enabling proactive measures to mitigate crises and foster more trustworthy online environments.

Figures (10)

• Figure 1: CrisisNews Dataset Creation Pipeline
• Figure 2: Yearly Distribution of News Articles in CrisisNews
• Figure 3: Distribution of Articles by News Publisher Category
• Figure 4: Distribution of Articles by News Publisher (Top 10)
• Figure 5: Stakeholder Relationship Matrix (Count)