With the debut of our new Television News Advertising Inventory Files, we're tremendously excited to see how scholars of advertising are able to use these datasets to better understand the language and visuals of television news advertising over the past decade.
Visual & Onscreen Textual Narratives
For researchers interested in the visual dimension of television advertising, the Video Time ASR-Aligned Dataset coupled with the Visual Global Entity Graph 2.0 is an extraordinarily unique resource. The VGEG 2.0 annotates every second of airtime across BBC News London, CNN, MSNBC and Fox News since January 1, 2020 and ABC, CBS and NBC evening news programming since July 2010, describing the objects and activities depicted within. The new ASR-aligned ad masking dataset catalogs each second of these annotations, determining whether it is news programming or advertising, allowing advertising scholars to hone in on just the visual annotations related to advertising.
The complete onscreen text of those advertisements is available as part of this dataset as well, allowing rich linguistic analysis of the text seen onscreen compared with the spoken narratives of advertising.
Spoken Narratives
For the roughly 30% of advertising airtime that is closed captioned, the Modified TTXT Dataset contains the full captioning of each, exactly as it appears in the captioning stream as provided by the commercial creator. This dataset offers an unparalleled opportunity to understand the spoken word language of television advertising over the past decade. While it covers only that 30% of advertising that is captioned (and there may be systematic patterns to which kinds of ads are captioned), this allows a rich analysis of the kinds of words used in advertising. Coupled with the visual narratives dataset above, which includes the full OCR'd onscreen text of each second of the ads, researchers can for the first time compare the spoken, written and visual language of advertising at scale. This dataset covers CNN, MSNBC and Fox News and San Francisco ABC, CBS, NBC and Fox affiliates (both evening news and 24/7 daily programming) back to 2010, business channels Bloomberg, CNBC and Fox Business back much of the past decade and Spanish-language channels Telemundo and Univision back a decade. This combination of general, business and Spanish-language news offers unparalleled opportunities for examining the language of advertising across both genres and languages.