WE’RE heading for yet another hung Parliament - and Julian Sturdy will lose his York Outer seat to Labour.

That’s the prediction being made today by a York-based ‘data mining’ company which has analysed more than 50 million Twitter conversations since the election began.

Steve Brewer of Text Mining Solutions predicts that the Conservatives are on course to lose almost 30 seats in the election compared to 2017 - leaving them with 289 seats, 37 short of a working majority.

Labour will be close behind, with 286 seats, he predicts - up 24 - while the Lib Dems will pick up just a single seat, and the Greens will lose their only MP, Caroline Lucas.

Perhaps most surprisingly, for York voters at least, Mr Brewer predcts that Labour’s Anna Perrett will win in York Outer, with the Lib Dems’ Keith Aspden in second place and incumbent Conservative MP Julian Sturdy trailing in third. He predicts that Labour’s Rachael Maskell will win with a healthy majority in York Central.

Mr Brewer admits that he’s a novice at calling elections. Since 2011, when he set up his business, he has built up an international client base. His sophisticated computer algorithm, which can ‘mine’ vast quantities of written or social media data for information, was used to help Coca Cola find an alternative natural sweetener for its drinks.

He has also done work for Body Shop, and is working with researchers at the University of Sheffield to analyse the carbon footprint of a traditional Christmas dinner.

But this is the first time he has turned his algorithm to the job of predicting an election result.

Nevertheless, he believes his system has generated a prediction that will be at least as reliable as the ‘notoriously inaccurate’ polling companies, who got the 2017 election so wrong.

The strength of his method is in the sheer quantity of data it can process, he says.

Since the general election campaign began, his algorithm has been processing something like 1.5 million tweets every day - more than 51 million altogether. It ‘mines’ tweets, hashtags and Twitter mentions for key words and phrases plus the names of politicians and political parties to find out which are being most talked about.

Further analysis is then done to reveal, on a scale of one to five, whether the people tweeting have strong negative or positive views towards the people or things they are tweeting about. From that, he says he can generate an accurate prediction of vote share.

Mr Brewer insists he has got no political axe to grind.

He says he’ll be voting on election day, like everyone else, but he’s not a member of any political party - and it’s in his interests to give as unbiased a verdict as he possibly can.

He says the whole point of going public is to showcase what his text mining algorithm can do.

Older people are less likely to use Twitter, however, so his sample base of Twitter users will be younger than the average voter, and may be more likely than average to vote Labour, Green or Lib Dem, he admits.

Mr Brewer said only time will tell if the predictions are correct - but that he couldn’t do much worse than the traditional pollsters in 2017.

He added that if he has got it wrong, he’ll be able to learn from the experience and improve his model for next time.

“I’m very excited to see what is going to happen!” he admitted. Aren’t we all?"