Twitter data 'predicts Boris Johnson victory' in London mayor election
BORIS Johnson will claim victory over Ken Livingstone in the London mayoral election, according to analyses based on data from Google searches, Twitter and Facebook.
The most recent opinion polls put Mr Johnson, who writes a column for The Daily Telegraph, four points ahead of his Labour rival, and the lead is seemingly reflected in online activity.
Google Analytics, a service that helps digital advertisers track web trends, shows that last week there were almost five times more searches for “Boris Johnson” than for “Ken Livingstone” via google.co.uk. Britons looked up the Conservative candidate online 11,629 times, compared to 2,337 searches for Mr Livingstone.
The digital marketing firm iProspect said search trends were also in Mr Johnson’s favour prior to his 2008 victory. Absolute numbers of searches from that time are not available, but Google shows that of the total for both candidates, 60 per cent were for “Boris Johnson”.
Neilson Hall, a spokesman for iProspect, said Mr Johnson’s relatively large lead in Google searches this time around - an 83 per cent share - reflected not only voters’ intentions but also a well-run digital campaign.
“Search has been overlooked as a key indicator of success in the run up to the election,” said Mr Hall. “It shows Boris has integrated better with the online community - you only have to look at the number of Twitter followers versus Ken."
"With Ken’s campaign it’s almost like digital is an afterthought, despite the proven impact it has on the overall result. For example just searching for his brand name and you can see negative websites."
The Tory lead in Google searches is repeated on social networks, according to analyses by iProspect and Lithium, another digital market firm, which both analysed posts on Facebook and Twitter, as well as forums and other web material.
iProspect said its “sentiment analysis” found 7 per cent more positive sentiment towards Mr Johnson than Mr Livingstone in the past month.
The analysis aimed to capture how often each candidate was discussed, how often individuals mentioned them, as a measure of “passion”, and the ratio of positive and negative language used.
Lithium said its social network sentiment analysis tools predicted 54 per cent of the vote for Mr Johnson, on a head-to-head basis with Mr Livingstone.
Dr Timothy Wu, Lithium’s chief scientist, said that using social network data to predict elections was in its infancy but had been shown to be accurate in US elections, including the recent Republican primaries.
“The important question is not ‘whether social media data can predict election outcome?’ It definitely can,” he said.
But, he said, social network analysis could only predict outcomes accurately up to two weeks before an election.
“Trying to predict anything beyond the predictive window of your data is, pretty much, useless,” said Dr Wu.
The bookmaker William Hill today shortened Mr Johnson’s odds from 1/5 favourite to win to 1/8. Ken Livingstone is on 9/2.