We are incredibly excited today to announce the debut of the relaunched Television News Explorer 3.0 in collaboration with the Internet Archive's Television News Archive. This public interest experiment allows journalists and scholars to keyword search the Television News Archive back to July 2009 for CNN, MSNBC/MSNOW and Fox News. You can search for words and phrases up to five words and use basic boolean searches like "(election OR elections OR voter OR voters OR voting OR voted)". Searches are exact matches, so you must specify all the words that should match such as "voter OR voters OR voting OR voted", etc.
For each search, you will get back a timeline showing how many times your keyword was mentioned in the spoken transcript of each selected channel each day and you can use the dropdowns at the top of the timeline to instead view it in weekly, monthly or yearly resolution and add rolling average smoothing to make the longitudinal trends more visible. You can also download the timeline as a publication-ready PNG image or a CSV to graph yourself. You can click-drag on the timeline to zoom to a narrower period, which will also cause the recent mentions list beneath the timeline to refresh.
Under the timeline you will see up to 75 of the most recent mentions of your keyword across the selected channels. If you zoom into the timeline, this pane will refresh to show the most recent 75 results as of the zoomed end date in the timeline.
When you first visit the TV News Explorer homepage, you'll also see links to the Visual Explorer and a list of Today's Media Trends briefings.
To explore what searching international coverage might reveal, we've also included IRINN and PressTV in this inaugural release, with IRINN being searchable both in its native Persian and English translation. All transcriptions and translations are either broadcaster-provided or are machine generated and may have errors or be incomplete. The Archive's coverage of a given channel may have outages and channels were added to the Archive at different times, so you will want to verify the trends you see, especially periods of highly reduced or no mentions of a given query.
We cannot wait to see what new kinds of findings you discover!
Please contact us with any errors you encounter.