EU-funded project claims correct Brexit prediction

Researchers on a project that received funding from Framework programme 7 to computationally analyse social media have said that they correctly predicted the outcome of the UK’s referendum on its membership of the EU.

Giuseppe Riccardi, an information scientist at the University of Trento in Italy and the coordinator of the Sensei project, which combines computational and human analysis, was quoted on the European Commission’s Cordis research project website as saying that the project correctly predicted “with very high accuracy” the referendum outcome of a majority of people voting to leave the EU.

The project analysed more than 6 million social media updates relating to the referendum to make its prediction, Riccardi said. According to the Cordis story, social media updates changed during the day of the referendum such that by the “late afternoon” the analysis indicated there had been a “dramatic swing” towards a vote to leave the EU. By comparison, pollsters continued to predict a majority vote to remain in the EU late into the evening.