How the Reveal Project aims to help journalists verify eyewitness media

Eyewitness media from social networks is now an integral part of many news stories, but the number of tools and platforms available can be overwhelming and make it harder for journalists to figure out which ones they should use and how in order to verify these materials.

The Reveal Project is an initiative co-funded by the European Commission, bringing together 11 European partners, including German broadcaster Deutsche Welle, technology companies and research institutes.

Their work began in 2013, focusing on the different components of eyewitness media and including case studies, tools and best practices. The research outcome will be showcased on the Reveal platform scheduled to launch in 2016.

"There's a lot of value in eyewitness media and it cannot be neglected, as it can significantly enhance and improve newsgathering and storytelling," Jochen Spangenberg, innovation manager for Reveal and Deutsche Welle, told Journalism.co.uk 

He said the current process of verifying this material is "quite a laborious one" and people might not see the value in the different tools and platforms available, so the project aims to simplify this approach.

Researching use cases

Reveal's work is divided into different components, as the project members work individually or together to analyse aspects from computer-assisted verification to the legal implications of scraping data and other obstacles faced by newsrooms dealing with UGC.

Spangenberg said the project also aims to "raise awareness and point to other resources available, so that people can have one platform they can use to tackle verification challenges".

Currently, the website provides weekly round-ups of discussions from around the web surrounding verification, but also findings of the ongoing research conducted by members.

For example, the IT Innovation Centre at the University of Southampton is developing ways to automate certain elements of verification that do not require human input, by analysing keywords and images.

Other research includes contributor analysis, for finding out whether or not an eyewitness from social media is reliable and what elements from their background can be gathered automatically.

A third component is multimedia analysis, performed by The Centre for Research and Technology Hellas (CERTH) in Greece, which is working on image verification solutions and comparing it to existing results from tools such as TinEye.

Spangenberg said these elements will be implemented on a level that can be "used in the day to day work environment of journalists" when the project concludes next year.

Research done by Reveal has three use cases that serve as a guide. The two journalism scenarios focus on event coverage by reporters and eyewitnesses, as well as gathering, organising and incorporating material from social networks into a final multimedia package.

There is also an enterprise scenario, which analyses content that is "residing in closed networks like businesses and enterprises in order to find people who have expert knowledge in particular areas".

From an editorial perspective, the work is done in close collaboration with Deutsche Welle's editorial teams and departments that use social media, looking at what challenges and requirements they face.

"Ideally, we will come up with a solution that really fits Deutsche Welle journalists in the first instance.

"If that works, it should also be deployable in other newsrooms that are facing similar challenges and produce something of value for people working with user-generated content," Spangenberg said.

Building new tools

The Athens Technology Center (ATC)'s Innovation Lab is the technical director behind the project and has been working with Deutsche Welle on a prototype called TruthNest, which will also be incorporated in the final version of the Reveal platform.

TruthNest is a web based platform that allows journalists to verify material posted on Twitter. When they log in with their Twitter account, users see the usual stream of tweets from their timeline, but with a few added elements.

The platform uses an algorithm that recognises names of people, locations and organisations mentioned in a tweet and groups them into categories.

There is also an option to create new streams, similarly to Tweetdeck – for example, a journalist can choose to monitor what an individual is posting about a certain topic, like Greece, or select the keyword to see only posts about that from everyone in their timeline.

Each tweet can be analysed, giving information across three categories: contributor, content and context.


Screenshot of the TruthNest demo, showing the different analysis option for a tweet

For contributors, which are eyewitnesses or sources, TruthNest gives the journalist information on a person's reputation, presence and influence on Twitter.

The tool also does sentiment analysis to find out how many negative, positive or neutral mentions the author has received on a given tweet or a range of recent posts, that can be further drilled down into.

For images, TruthNest uses the TinEye API to show when and where a picture has been posted on other websites before it reached the social network.

The context category allows journalists to see who else is sharing information about a given topic by grouping together tweets containing similar keywords from around the world. There is also a proximity check, which compares a tweet's geotag with the location the user claims to be posting from.

"It's a quick way to gather many clues around a tweet," said Nikos Sarris, head of operations at the ATC Innovation Lab, "but it is up to the journalist to decide whether it gives valuable insight or not."

Once TruthNest is fully developed, Sarris is hoping to monetise it on a subscription basis, according to a news organisation's size and the number of searches they do per day using the platform.

• If you are a journalist dealing with social media analysis and UGC verification and would like to provide feedback or test prototypes and components as they emerge from Reveal, get in touch with Nikos Sarris (@nikossarris) or Jochen Spangenberg (@jospang).

• Find out more about the challenges of verifying eyewitness media from social networks and the tools available for journalists at news:rewired 'in focus', a half-day event taking place on 21 October at Reuters in Canary Wharf, London.

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).