Last week, the Internet Archive released an interactive map visualization of its Television News Archive, a trove of 508,000 news programs collected over 4 years from national U.S. networks and stations in San Francisco and Washington D.C. The visualization tracks approximately 400,000 hours of coverage, with redder dots indicating more mentions and yellower indicating fewer. Clicking on a location brings up a graphical overview of its coverage on U.S. TV news.
With such a massive dataset for source material, some errors are inevitable. But errors aside, the prototype is a fascinating proof of concept for understanding the geography of TV news coverage.
The Internet Archive collaborated with Georgetown University computer science fellow Kalev Leetaru to automatically scan the closed captioning of each broadcast to capture and map each location mentioned. Dr. Leetaru is one of the co-creators of the Global Database of Events, Language and Tone, a catalog of events drawn from international news stories that has also been mapped in interactive time-series.