METHODS

R2S uses large-scale scraping, automated transcript analysis, and publicly available data to study internet culture empirically. It combines quantitative analysis at scale with qualitative close reading, corroborated by on-the-ground research and interviews.

Each project follows four stages:

• Collection: scraping platforms, pulling public datasets, building corpora
• Analysis: automated coding with large language models for sentiment, worldview, and ideological framing across full datasets, paired with qualitative close reading of representative cases to check patterns and surface what the numbers miss
• Verification: on-the-ground research, studies, and interviews to corroborate the data findings
• Visualization: translating patterns into reports designed for legibility

Methods and caveats are documented in each report.