A team from Deutsche Welle developed a data mining and visualization prototype at the event, helping editors better understand content and topics.
Over the last year Facebook has started initiatives to build stronger ties with media and publishing organizations, e.g. the “Facebook Journalism Project“. The goal here is to gather both journalists and developers to collaboratively build products and services related to news content, news distribution, and digital storytelling. Ideally, the outcome benefits everyone: Users, journalists and, of course, Facebook, too.
DW at Hamburg Hackathon
As a part of their project, Facebook recently organized several “hackathons” where media and tech people were asked to work on specific ideas for one day and come up with prototypes. One of those events took place in Hamburg in July. DW attended with a team of four: Tobias Doll (Projects & Development), Alexander Plaum (Research & Cooperation Projects), Okay Suna (Projects & Development), and Andres Tabares (Planning & Evaluation). They were joined by hackers from Gruner + Jahr and We-Build.City. This combination proved to be a good one, as the jury has recently announced that the DW project won the 2nd place.
To improve services, look at the data
Development of the winning entry started with a question: How can we help improve DW News (= Deutsche Welle’s English language news service) on Facebook? Would it help if editors had more insights and information?
Witin a couple of minutes, the idea of “DW Mining” came up: Parse DW News posts on Facebook and (automatically) visualize the findings in shareable infographics. To this end, the team set up a couple of tools and started to analyze the Facebook data, always led by questions pertinent to the editors in the newsroom, e.g.:
Which cities have been in the in the spotlight recently?
Or: Is there a region that has been neglected in reporting, even though it’s potentially quite interesting?
Future versions of the tools will probably feature semantic connections as well answering questions like:
Do stories that involve politicians X, Y or Z usually lead to positive or negative reactions?
DW Data is a tool combining data mining and data visualization concepts. The technical workflow behind it basically looks like this:
DW Mining Video
The DW team also documented their approach (and the prototype) in a short video:
Other winners of the Facebook Hackathon
The “gold medal” was secured by a Berliner Morgenpost team and their Social-Recommendation-Widget, 3rd place went to project F.i.d.T. (“finding fresh voices on Facebook”). All top-ranked projects will be revisited by Facebook and possibly further delevoped by the media organizations involved.