Case Study: Thomson Reuters separates real news from fake news on Twitter in 40 ms with Cloudera

Thomson Reuters separating real news from fake news on Twitter in 40 milliseconds

Thomson Reuters, a global leader in news and information for financial, legal, tax, accounting and media markets, faced the challenge of rapidly sifting millions of tweets to separate verified news from opinion and misinformation so journalists and clients could act on trusted information in real time.

They built Reuters Tracer, a machine-learning-driven service (powered by Cloudera) that analyzes ~13 million tweets daily, evaluates hundreds of features, assigns a newsworthiness/credibility score, and captures events in under 40 milliseconds. The system has uncovered major events ahead of other outlets and lets journalists focus on higher-value reporting while alerting customers to market-moving developments immediately.

Thomson Reuters

Khalid Al-Kofahi

Head, Corporate Research & Development