We stand with Ukraine

Tweeting Brexit: What cognitive software revealed in analysis of Brexit tweets

Expert.ai Team - 6 April 2017

The now infamous EU Referendum in the UK last June and the resulting decision to leave the European Union impacts not only Europe, but the entire world. While the process of the UK leaving Europe is just getting started, the main question “Should the United Kingdom remain a member of the European Union or leave the European Union?” is still open for discussion.

At Expert System, we had the chance to foresee the results of the referendum using cognitive text analytics. Using our Cogito cognitive technology, we collaborated with the University of Aberdeen to analyze more than 55,000 tweets posted from June 7 to 15, 2016 in England, Northern Ireland, Scotland and Wales. We also analyzed a sample of 5,000 tweets collected from June 20 to June 21 to discover the most discussed Brexit topics on Twitter and highlight the differences between the first period of the analysis and the days closer to the referendum.

[su_button url=”https://devexpertai.wpengine.com/expert-system-university-aberdeen-find-64-75-tweets-britain-inclined-leave-european-union/” target=”blank” style=”flat” background=”#d40000″ color=”#ffffff” size=”4″ center=”yes” radius=”auto” icon=”icon:arrow-circle-down” icon_color=”0000″ text_shadow=”0px 0px 0px #000000″]Download the report[/su_button]

Brexit sentiment: Leave or Stay?

To understand the voting intentions of users on Twitter, we used natural language processing and text analysis to quickly and accurately understand text as a human would by establishing precise connections between concepts.

Our Brexit Twitter analysis found that jobs, immigration and government issues were the most discussed topics in all four countries as they were included in all of the tweets analyzed. In thehe days before the vote, immigration was the focus of most of the debate on Twitter in the countries considered for the analysis.

Brexit sentiment analysis on Twitter showed how sentiment changed over the course of the analysis period. The gap between Leave vs. Stay grew in both England and Scotland toward the end of the first period of analysis. In Ireland, both sides showed a decrease at the end of the first period and then an increase; trends in Wales were more stationary during the first period of analysis.

The majority of tweets from each of the countries showed a desire to leave the EU. Expert System and the University of Aberdeen found that 64.75% of tweets from Britain showed an inclination to leave the European Union. The majority was higher in England and Scotland and slightly lower in Northern Ireland and Wales.

At Expert System, we use our technology to analyze information streams using cognitive computing technology that can understand content, context and sentiment as expressed in text content on any platform. This ability to understand and listen can be used to predict events, identify trends and tipping points, address challenges, identify business opportunities or detect threats.

If you are interested in learning more, including additional data analysis of “What people said on Twitter about Brexitdownload the report here.