Social Media in the Newsroom workshop at ICWSM – round up

The Social Media in the Newsroom workshop took place on the 17 June 2016 in Cologne, as part of the ICWSM Conference. The workshop was aimed at the intersection of social media and journalism, as a subset of Computational and Data Journalism, with a focus of the ever growing ability of social media for capturing and announcing breaking news. The workshop participants were keen on furthering the development of tools to facilitate access to social media for news and media professionals, as well as on gaining more insights into social media’s benefits from a journalistic perspective.

The workshop kicked off with an insightful keynote speech, delivered by Wilfried Runde, the Head of Innovation Projects at Deutsche Welle. In his talk, Wilfried outlined the main challenges that newsrooms need to deal with when gathering news from social media. The talk largely focused on social media verification practices at Deutsche Welle, which he illustrated with numerous examples of fake images and videos that made it easier for everyone to understand the challenges. Wilfried also showcased some of the work they have conducted within the Reveal EC-funded project, as well as some of the forthcoming plans for a new project that focuses on video verification, InVID. The audience was curious about these verification challenges, and raised a number of interesting questions, including the ability to perform thorough verification tasks at scale.

Screen Shot 2016-07-01 at 14.28.52

After the keynote talk, the workshop featured presentations from authors of 5 accepted full papers. Markos Zampoglou presented the first paper, which was an appropriate follow-up to the keynote talk, as it also focused on verification, in this case looking into the development of automated tools. His presentation included a demo that resulted from the research, called ‘Media Verification Assistant’. It allows its users to upload a picture, producing an enhanced output that includes a likely veracity value for the picture. The second paper was presented by Konstantina Papanikolaou, who introduced PALOMAR, a tool that uses named entity recognition and event extraction techniques for detecting events, aggregating both social media and news media in Greek. For the third paper, Mossaab Bagdouri presented research on using a classifier to determine what types of questions Arab journalists make on Twitter. They developed a scheme of 7 types of questions, and annotated a collection of Arabic tweets from journalists asking questions. Using a supervised classifier, Mossaab showed that they can accurately classify questions by type, which they plan to extend in future work with larger collections of annotated tweets. The fourth paper was presented by Laura Tolosi, who has been studying rumour detection techniques within the PHEME EC-funded project. She presented a study of features that characterise rumours across different events, suggesting those that can be exploited to develop an accurate rumour detection system. To conclude, Cornelius Puschmann presented his paper looking at the news sources of Islamophobic groups on Twitter. His research revealed findings with implications both for the study of mass media audiences through the lens of social media, and for research on the public sphere and its possible fragmentation in online discourse.

Screen Shot 2016-07-01 at 14.29.07

Interestingly, the audience actively engaged in all the presentations, and raised questions that ignited discussion. While the half-day workshop came to an end quickly, the good turnout and active discussions posited a collective interest in following up with further workshops and activities.

The workshop was organised and moderated by Bahareh Heravi and Arkaitz Zubiaga.

Leave a Reply

Your email address will not be published. Required fields are marked *

You may use these HTML tags and attributes: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <s> <strike> <strong>