Television News Spoken Word Neural Entity Graph (TV News GEG) Now Updating Daily

We're tremendously excited to announce that the television news spoken word neural entity graph announced last month has now been updated through present and will update each morning! The Television News Global Entity Graph (GEG) processes the closed captioning transcript of all BBC News, CNN, MSNBC and Fox News broadcasts and the ABC, CBS and NBC evening news broadcasts through Google's Cloud Natural Language API and constructs a minute-by-minute chronology of the topics and things they discuss.

The downloadable JSON chronologies are updated each morning to the previous 24 hours, while the BigQuery table is updated to 7 days prior each morning, similar to the Television News Ngrams dataset.

We've also added a new "lang" field that contains the Google-provided language code of the article used by the API. This field is present in the JSON files from December 5, 2019 onward and was retroactively added to all records in the BigQuery table. If absent, assume a value of "en" for English. This field was added to support future work as we expand our neural analyses beyond English television news.

We're tremendously excited to see what you're able to do with this incredible new deep learning-powered lens!