Al Jazeera creates interactive video transcripts of debates

The four US presidential debates maybe over but Al Jazeera English now offers a way for readers to view, explore and share sections of each debate through interactive video transcripts.

Each word in the videos of the debates is linked to that word in the transcript, enabling the audience to search by keywords and find the exact point in the video where that word is spoken. Users can then choose to share that point in the footage via social media.

The transcripts also allow users to search for a keyword and see a pie chart and graph showing how many times the word was used by each candidate and at what points of the debate.

You can try searching for a word such as 'economy' or 'jobs' (or 'horse' or 'bayonet') in this video of the fourth debate (or in the first, second and third of the videos).



The interactive video transcripts were created by Mark Boas, one of five winners of a 2011 Knight-Mozilla OpenNews fellowship who is funded to work with Al Jazeera and "bridge the gap between technology and news", and Al Jazeera colleague Mohammed Haddad.

Boas, who has been working with Al Jazeera remotely for the past 10 months, and Haddad "compressed about a month's work into two-weeks", Boas explains in a blog post in which he describes the process - and the limitations of the approach.

The blog post was written after the first debate and Boas and Haddad have since been iterating and refining the process.

Boas told Journalism.co.uk that he used jPlayer, a media library for audio and video, and Popcorn, a javascript library for the interactive video, "taking it a step further" by using a charting library called D3js to create the animated pie charts and graphs.

Early in his placement at Al Jazeera Boas did a piece of work based a video called The fight for Amazonia, which pulled text content from a Google Doc and linked words to the exact points in the films creating "contextual video" (you can read about  that project here and see an example here).

"We've now taken that a lot further," he said. "There are now four debates on Al Jazeera that you could call interactive that have this kind of hyperaudio structure to them."

He added that users might decide to "highlight an excerpt that they find interesting a share that through a tweet or through Facebook".

Boas explained that they used a third-party service to transcribe the text. "One of the problems is that we had to wait for the transcript to become available," he explained, with that being problematic in that most of the interest in the debates occurred during or soon after the events took place. The first transcript was added 18 hours after the debate with the time being reduced for subsequent debates.

The Guardian US interactive team has also produced an interactive with on the text from the debates.

Free daily newsletter

If you like our news and feature articles, you can sign up to receive our free daily (Mon-Fri) email newsletter (mobile friendly).