Skip to content
News agency

The making of Reuters News Tracer

Published on

How the need to filter noise from social media to deliver trusted news worldwide led to a versatile tool for social media events monitoring.

At Thomson Reuters, we take our Trust Principles very seriously. The preservation of independence, integrity and freedom from bias is fundamental to the way we do business, so it’s only natural that our internal teams would establish a way to crowdsource news notifications from the people who see events first via the most unlikely of sources: Twitter.

Sameena Shah, our director of research for technology development, and Reginald Chua, the executive editor for editorial operations in the data and innovation division of Reuters, led the way in building Reuters News Tracer from the ground up in an innovative intrapreneurship.


View the 2016 Annual Report: The Partnership Equation to see highlights from our strategic partnerships with clients, startups and universities over the last year.


What is Reuters News Tracer?

Reginald Chua: Reuters News Tracer is a capability that we’ve developed with our R&D team that finds events that are breaking on Twitter. It assigns them a newsworthiness score so you can focus on the things that are important, and the real magic of it is that it then gives a confidence score about how likely it is that those events are true. This is really critical because the landscape of news has changed dramatically.

Screenshots from Reuters News Tracer, a Social media listening tool, show event cluster features, including newsworthiness, veracity, topic clouds and other metadata in real time social media monitoring via machine learning.
Reuters News Tracer event cluster features include newsworthiness and veracity scores, topic clouds and other metadata in real time. Source: Recorded Reuters News Tracer demo.

How was this social monitoring tool conceived?

Sameena Shah: Over the last almost one decade we have seen an increasing amount of news events unraveling first on social media. In some ways we can say that social media has actually disrupted the news industry and journalists do need to tap into that information and signals that are available in social media. So the mission really was to tap into the social media content to figure out what the news events really are and present them in a timely manner as soon as possible. The goal was to effectively get rid of all the spam, the noise and only harness the signal or the news events that are important for our journalists and eventually our customers.

Chua: So when we sat down with the R&D people it was really a question of: How much can we do? How can we narrow the scope of this so we can be successful? And to their credit, our development team basically said: “Let’s make this big. Let’s try to solve much more universally than solve for individual small uses.”

Shah: We interviewed numerous journalists all across the globe from various specializations and really listened to how they gather information; how they use social media to assess whether something is newsworthy or not; and how do they go about verifying various types of different stories. And once we understood that we created the algorithms and tools that could mimic just that.

Is Reuters News Tracer good at spotting news?

Chua: One thing we found is that it’s been very good at finding certain types of events much more quickly than many mainstream news organizations are able to do. It was ahead on the Brussels Airport bombing by several minutes. It was ahead on the Chelsea bomb by again several minutes.

Since we started keeping analytical records about a year ago, Reuters News Tracer has beaten global news outlets in breaking over 50 major news stories. This has given our Reuters journalists anywhere from an 8- to 60-minute head start.

So it’s done very well and we can see that as a tip service for news organizations, and for us particularly, obviously, it’s been very helpful. It lets us start getting on a story much more quickly.

What makes this innovation partnership special?

Shah: It was really a very, very cohesive partnership in which technology brought in the technical expertise, the scientific rigor that could actually build those algorithms. And the human expertise that we were wanting to mimic, that was brought in from the Reuters journalists.

Chua: What’s great about this partnership is that to be able to take a fire hose of information that comes in at 140 characters and, with not much more than that, to be able to say: “This is newsworthy, you should look at it.” And then to say: “This may well be true.” That, I think, is a really interesting place to be.


Learn more

Check out the process and methodology behind Reuters News Tracer in our 2016 Annual Report and view our recorded demo to see Tracer in action.

Follow @Reuters for verified, breaking news.


Join the conversation

How do you get your news? Let us know in the comments below.

  • Facebook
  • Twitter
  • Linkedin
  • Google+
  • Email

More answers