Document Type
Poster
Publication Date
Spring 2026
Abstract
This project examines the evolution of digital journalism through a longitudinal NLP analysis of BuzzFeed headlines from 2012 to 2025. Using an NLP pipeline, the study applied cosine similarity, sentiment analysis, frequency modeling, and topic modeling to measure changes in language and focus over time. The findings reveal a decline in the listicle-heavy style that defined BuzzFeed’s early success and a rise in shopping and affiliate-driven headlines built around consumer recommendations and curated internet reactions. In 2016, positive language and numbered list formats (“17,” “19,” “best”) dominated engagement, while in 2025, terms such as “products” became the most prominent. Topic frequency analysis further showed declines in clickbait framing words like “reasons” and “ways,” alongside sharp increases in shopping, home, TV, and quiz-related content. Taken together, these linguistic and topical shifts suggest that BuzzFeed’s business model has moved away from virality and social sharing toward direct monetization through commerce publishing. More broadly, the study argues that this reflects a wider collapse of viral digital media, where platforms that once prioritized relatability and cultural identity are now using that same voice to drive consumer conversion rather than community engagement.
Recommended Citation
Eisenbeis, Gwen, "What NLP Tells Us About the Future of Viral Journalism: A Case Study on BuzzFeed’s Strategic Pivot (2016–2025)" (2026). IPHS 484: Senior Seminar. Paper 47.
https://digital.kenyon.edu/dh_iphs_ss/47
Creative Commons License
This work is licensed under a Creative Commons Attribution 4.0 License.