Visualizing the Who and What of Cable TV News
What would you ask if you could analyze the news yourself?
Each day, cable TV news networks determine what information millions of Americans receive. They also set the context and tone of the information presented. Editorial decisions, about who appears on cable TV news and what they talk about, shape public opinion and culture. Many newsrooms and monitoring organizations, like Media Matters, American University’s Sunday Morning Monitor Archive and NPR’s On The Media, routinely audit the content of news broadcasts, but these efforts involve the tedious and time-consuming work of manually counting who and what is on air.
To increase transparency around daily editorial choices, the Stanford Cable TV News Analyzer uses modern AI techniques to automatically measure who is on the news and what they talk about. We have applied this tool to over a decade of nearly 24–7 cable TV news broadcasts by CNN, FOX News, and MSNBC — over 270,000 hours since Jan 1, 2010.
The WHO of cable television news
The Stanford Cable TV News Analyzer uses automatic face identification to measure the amount of screen time that public figures receive on cable TV news broadcasts. For example, the chart below compares the screen time of President Donald Trump and Former Vice President Joe Biden over the last decade.
(Click the link in each of the following captions to view and edit the interactive graphs.)
Trump’s screen time jumps dramatically with the announcement of his presidential candidacy in mid 2015. This summer so far, with the 2020 Presidential Election race in full swing, Trump’s face remains on screen about two to three times more than Biden’s.
You can dive deeper into the data by filtering results by cable news channels, programs, and time of day. For example, Trump has generally received the most screen time by cable news channel on MSNBC (orange) for most of his presidency, and on CNN (teal) he has had the least amount of screen time. The gap between MSNBC (orange) and Fox News (red) closed in late 2019. These two cable news channels have had a similar pattern of screen time ever since.
In addition to plotting the screen time of specific individuals, the Stanford Cable TV News Analyzer can also plot screen time of groups, such as male- and female-presenting individuals, or major news presenters (hosts, on-air staff, etc.). Future posts will describe analyses using these features, but for now we encourage you to try the tool yourself. For example, examine how much screen time leading female candidates received during the 2020 Democratic primary season.
The WHAT of cable television news
In addition to measuring face screen time, the Stanford Cable TV News Analyzer also makes it easy to view trends in how often words or phrases occur in news program transcripts. This enables analysis of the stories and topics are covered on these cable news channels. For instance, the graph below compares how often the countries Russia, China, North Korea, Iran, and France are mentioned on cable news.
These results show that there is a noticeable change in the pattern of coverage of international news events after the start of the Trump administration. Russia, China and North Korea have been mentioned more often since he has been in office.
You can also write queries that combine filters for who appears on the screen and what is said in the transcript. According to some news analysts, one of Trump’s most common catchphrases during the course of the 2016 election was the phrase “believe me”.
We find that during Trump’s 2016 election campaign and through the first year of his presidency, Trump’s face was on screen almost half the time the phrase “believe me” was said on the news. Combining filters for who appears on the screen and what is said in the transcript can provide quantitative data insights on the word choices Trump, or other public figures, use.
Watch source video
The Stanford TV News Analyzer is a measurement tool. It cannot determine whether the motivation behind editorial decisions is the result of the news cycle, editorial choices or ratings goals. However it can generate quantitative data about the content of the news, and then facilitate close examination of the associated video. At any point when using the tool, clicking on the graph will show the videos corresponding to the exact moment where an individual appeared or a word was mentioned.
How does it work?
The Stanford TV News Analyzer has applied deep-learning-based image and audio analysis processing techniques to nearly a decade of 24–7 broadcasts from Fox News, CNN, and MSNBC going back to January 1, 2010. That’s over 270,000 hours of video updated daily. Computer vision is used to detect faces, identify public figures, and estimate characteristics such as gender to examine news coverage patterns. To facilitate topic analysis the transcripts are time-aligned with video content, and compared across dates, times of day and programs. To learn more about the automatic video annotation process as well as its accuracy, please see our methodology page and frequently asked questions.
What do we hope to accomplish?
Cable news is still one of the largest sources of news information and the Stanford TV News Analyzer presents a powerful new way to understand the representation of people and events on cable news. How long are certain people on the screen? How often are certain words mentioned? What will you find when you compare these measurements across time, channel, and programs?
Try it out yourself
Right now you can use the Stanford TV News Analyzer yourself to take a look at the 2020 American election campaign, or count the screen time of politicians, newsmakers, and experts you might be curious about. In 2020, as a result of the death of George Floyd, journalism and media outlets were questioned about their objectivity. Consequently, major management changes and diversity initiatives have been issued at many media publications, including CNN, The New York Times, The Philadelphia Inquirer, Bon Appétit, and Refinery29. This tool can be used to measure coverage of related stories. In future weeks, we’ll post more about these topics
Examine how the world’s biggest stories are being covered and how your favorite cable news network compares to the competition. We invite you to our Getting Started page and to learn more about how to measure the contents of cable news on your own.
Our Team
Stanford University Student Research Assistants: James Hong (student lead), Jacob Ritchie, Jeremy Barenholtz, Will Crichton, Daniel Fu, Ben Hannel, Michaela Murray, Xinwei Yao, Haotian Zhang
Maneesh Agrawala: Forest Baskett Professor Computer Science and Director of the Brown Institute for Media Innovation at Stanford University
Kayvon Fatahalian: Assistant Professor of Computer Science, Stanford University
Geraldine Moriba: Journalist, Documentary Filmmaker, Broadcast News Executive, Stanford University Brown Institute for Media Innovation Fellow and Research Scientist, and former John S. Knight Journalism Fellow
Acknowledgments: The Stanford Cable TV News Analyzer is a collaboration between researchers at Stanford University’s Computer Science Department, the Brown Institute for Media Innovation, and the Internet Archive.
