QM entrepreneurs visit Downing Street for high-level technology talks
The founders of a Queen Mary, University of London spin-out company joined the country’s leading technology experts at 10 Downing Street last week to discuss start-ups, sentiment detection and supercomputers.
17 January 2012
Dr Stuart Battersby and Dr Matthew Purver of the Interaction, Media and Communication group at Queen Mary’s School of Electronic Engineering and Computer Science, founded commercial venture Chatterbox Analytics in 2010 to build tools for analysing social media.
As owners of a growing technology start-up, they were invited to roundtable discussions with David Willetts MP, Minister of State for Universities and Science, and Justin Rattner, Chief Technology Officer of computing corporation Intel. Talks were focussed on boosting UK output, jobs and competitiveness through technology.
The event on Tuesday 10 January saw a small number of key players from across academia, government and industry participate in talks that focused on how the UK’s technology sector can expand and compete on the global stage.
Dr Stuart Battersby was asked to comment at the event on Chatterbox’s collaboration with Intel Corporation. In 2011, Intel showed its support for technology start-ups in east London, including Chatterbox Analytics, by providing one of its supercomputers for use, free of charge.
Chatterbox, a business that provides tools for automatically analysing patterns of conversation, was one of the first companies to trial Intel’s supercomputer. They used it to build their sentiment detection software, Chatterbox Sentiment Detection API. Intel has allowed new and growing businesses to use the technology to make sure their websites and digital services still function if there is an unexpected increase in use.
“Chatterbox Analytics used Intel's high performance computing platform to refine our innovative engine for detecting the sentiment of short social texts such as tweets or Facebook statuses,” says Dr Battersby. “This enabled us to experiment with varying parameters and use millions of tweets to train the computer system to recognise the slang and niche language used on Twitter.”
“The experiments Chatterbox carried out using Intel's supercomputer increased the accuracy of our sentiment detection engine from 85 per cent up to 90 per cent, representing a 33.3 per cent decrease in error rates. Access to Intel's resources have accelerated Chatterbox's development, bringing our sentiment analysis API to market ahead of schedule and thus allowing us to gain traction in the market."
Chatterbox Analytics are currently looking for consumer facing brands to help trial a beta version of their Twitter-analysing software. Their unique social media analytics service allows organisations to see who is driving discussions about brand-related issues through Twitter. Statistics on the content of messages can determine not only which messages are positive or negative, but how this develops as conversations progress, giving organisations vital knowledge with which to understand the context of conversations and then interact with an audience.
They are looking to expand and make use of other research from Queen Mary’s School of Electronic Engineering and Computer Science. As the software develops, Chatterbox will be extending the service to cover Facebook and Chinese microblogging website Weibo.
For further information on Chatterbox’s beta trial, visit;
For further information on the Chatterbox Sentiment Detection API, visit;
For media information, contact:Neha Okhandiar
Public Relations Manager
Queen Mary University of London