What is it about?
Social networks are packed with posts, but only a few are real news. We collected tweets and Facebook posts, crowd-tagged their news value, then trained lightweight machine-learning models on cues like message type, comment count and sentiment. The system now spots journalist-relevant content with about 80 % accuracy, helping reporters surface stories fast.
Featured Image
Photo by Hans-Peter Gauster on Unsplash
Why is it important?
Journalists risk overlooking breaking stories buried in endless timelines. Our model instantly sifts millions of posts and flags the few with real public-interest value, letting reporters spot and verify news sooner, broaden coverage, and cut the time and cost of manual monitoring.
Perspectives
Developing this tool let me "sit beside" working reporters and feel the stress of endless scrolling. Turning their gut sense of “this is news” into fast, transparent code was both frustrating and exhilarating. I’m proud that the final model is light enough for small newsrooms to run and clear enough for journalists to trust. I hope that it gives them back precious minutes for verifying facts and crafting the story, not sifting through noise.
Prof Alvaro Figueira
University of Porto
Read the Original
This page is a summary of: Detecting Journalistic Relevance on Social Media, July 2017, ACM (Association for Computing Machinery),
DOI: 10.1145/3110025.3122120.
You can read the full text:
Contributors
The following have contributed to this page
