Multi-Platform Social Media Discourse on Brand Activism (2020–2023)

The study compiles a multi-platform corpus of social-media discourse on brand activism spanning January 2020–December 2023, aggregating ~2.3 million posts/comments from Twitter/X (via the Academic Research API), YouTube comments (YouTube Data API v3), and Reddit discussions (PRAW from relevant subreddits), supplemented with Google Trends query volumes for issue/brand terms. Each record includes text content, engagement metrics, user/network metadata, and inter-user interaction links; the corpus is cleaned (deduplication, spam filtering, language detection, tokenization/lemmatization with sentiment-bearing cues preserved) and converted into graph structures (≈847k nodes, ≈3.2M edges) for GNN-based sentiment diffusion modeling, stance detection across ideological segments, and temporal anomaly analysis.