[ddj] new research tool for searching two million hours of television news

kalev leetaru kalev.leetaru5 at gmail.com
Mon Jan 9 17:32:47 UTC 2017


Based on great feedback from many of you, a new version of the system was
released over this past weekend, which now includes thumbnails and snippet
text in both the interface and JSON output, which can be used for things
like top terms word clouds or selecting expansion or refinement search
terms based on contextual keywords. The system also now computes a daily
ngram table over national US news coverage and differences those each day
to compute a set of top trending phrases each day to give you a sense of
the stories that are dominating the news cycle.

http://blog.gdeltproject.org/television-explore-interactive-time-zooming/
http://blog.gdeltproject.org/television-explorer-now-includes-bbc-news/
http://blog.gdeltproject.org/introducing-television-explorer-top-terms-word-cloud/
http://blog.gdeltproject.org/introduction-new-television-explorer-top-clips/
http://blog.gdeltproject.org/television-explorer-top-trending-topics/

Email me directly if you have any questions or suggestions!

Kalev

On Wed, Dec 21, 2016 at 10:01 AM, kalev leetaru <kalev.leetaru5 at gmail.com>
wrote:

> I thought many of you would find of great interest my latest collaboration
> with the Internet Archive - this time to create a new research
> visualization tool that allows you to visualize two million hours of
> television news programming spanning the last 7 years.
>
> You can specify any keyword and/or context keywords (allowing you to run a
> "near" search like "clinton NEAR email) and create a timeline charting by
> day how it has appeared in American television news over time and what
> networks focus the most on it.
>
> Unlike the Archive's primary Television News interface, which returns
> results at the level of an hour or half-hour "show," the interface here
> reaches inside of those six years of programming and breaks the more than
> one million shows into individual sentences and counts how many of those
> sentences contain the given keyword. Thus, instead of reporting that CNN
> had 24 hour-long shows yesterday that mentioned Donald Trump one or more
> times, this tool counts how many sentences uttered on CNN yesterday
> mentioned his name - a vastly more accurate metric for assessing media
> attention.
>
> CSV and JSON output are also available to make it easy to import the
> timeline into the analytic package of your choice like R or simply Excel
> for further analysis. More features will be coming over the next few months.
>
>
> http://blog.archive.org/2016/12/20/new-research-tool-for-
> visualizing-two-million-hours-of-television-news/
>
> http://television.gdeltproject.org/cgi-bin/iatv_ftxtsearch/iatv_ftxtsearch
>
>
> Kalev
> http://kalevleetaru.com/
> http://blog.gdeltproject.org/
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.okfn.org/pipermail/data-driven-journalism/attachments/20170109/51da3939/attachment-0002.html>


More information about the data-driven-journalism mailing list