UK Sales: +44(0)20 3048 3701  Canada Sales: +1 (647) 239-9872

Predicting the EU referendum using Big data

big dataTomorrow is the big day when all the talk will end and we will vote on whether the UK remains part of the EU or not. Are we clear on who the polls currently have leading the race? Can big data help us understand the likely outcome of the referendum?

Over the last few months most polls have had it nip and tuck between remain and leave with an undecided of significance that could swing it either way.

One of the interns here at Invennt, Dr Xuxin Mao has developed a big data framework, namely the topic Retrieved, Uncovered and Structurally Tested (TRUST) methodology.

This methodology provides a robust capability for forecasting through considering a number of different data sets and combines it with the kind of dynamic modelling that has been used to improve the forecasts of exchange rates, and other economic variables.

Recently the framework correctly predicted the results of the Scottish referendum and today is predicting a win for the remain camp.

Through supplementing polling numbers with further text mining and other data freely available to us, Xuxin predicts a narrow but comfortable win for the remain camp with the economy and not imigration being the most important factor in the conversation.

One of the more interesting finds is the negative impact David Cameron is having on the campaign!

More on the prediction can be found here:

If you are interested in further exploring the results or understanding how we can utilise big data analytics to predict future market trends in the construction industry please get in touch at or 07961071166

Ben Pritchard

Ben joined invennt from Magnox Ltd where as a Framework manager in the Nuclear Decommissioning sector he led the procurement and commercial management of a range of frameworks and projects covering demolition, new construction, refurbishment and retrieval, processing and conditioning of waste.

Leave a Reply

Your email address will not be published. Required fields are marked *