Trend Detection Using Television News Ngrams

The Television News Ngram (TV-NGRAM) dataset can be used to assess a wide range of topical, narrative and emotional latent undercurrents of television news programming. It can also be used to automatically detect breaking news and emerging narratives.

For example, an algorithm that analyzed the TV-NGRAM bigram dataset each day would detect the novel phrase "hurricane dorian" emerging for the first time on August 26, 2019 with a single mention, followed by 3 mentions the following day and then a surge to 34 mentions on the August 28th and 182 mentions on August 29th.

The same algorithm could detect that on September 7th and 8th coverage began sharply declining, suggesting the story was fading from prominence.

Trend analysis of the television news ngram dataset offers a tremendous number of opportunities for tracking the emergence and decline of news narratives and automated detection of stories, especially the emergence of subtle low-bore narratives and stories.