Here’s the story of how the recent elections in the Tower Hamlets London’s Borough went in the Digital World and how Websays could predict who was going to win based on a simple POPULARITY FORMULA = TOTAL MENTIONS – NEGATIVE MENTIONS.
It seems that both John Biggs and Peter Golds choose the same day for the campaign boost: the 5th of June. But although Peter Golds generated 86 tweets and retweets that very day, John Biggs got a pretty good amount of engagement with only 32 tweets and retweets.
As clearly seen in this graphic, the corporate channels worked pretty well to engage with their audiences, normally the external conversation is bigger than what we call “the response to corporate conversation”. John Biggs for instance got a 79% and Rabina Khan a 69% of the whole conversation as a response to their corporate conversation. Peter Golds got a bigger 83%, but generated double amount of corporate conversation.
Labour 'campaigners' too busy taking pictures - gets heated!
This youtube video was the piece of content with more hits, 3,000 by the elections day.
Marias
This Facebook post by the “Mothers Against Radical Islam And Sharia” had the highest impact (54 likes, 11 comments and 121 shares by election day).
@YvetteCooperMP
This was the Tweet with the most impact referred to the elections. The parties kept on tweeting about other matters during the campaign which had much more impact in the public opinion.
@SadiqKhan
This positive Tweet had the most impact.
5Pillars
This Facebook post was the most negative – with 79 likes and 6 shares by election day.
And to finish this short report, here we show the sentiment summary of the two main candidates which show how citizen’s feelings expressed in social networks are a valuable indicator in voting prediction.
With 25,594 mentions gathered from the Internet within 30 days prior to the elections, mostly from Twitter, followed by YouTube. The results show how this controversial elections went and how at Websays tracked the popularity. Although the amount of mentions were small for full analysis, we could demonstrate that we are closer than traditional surveys. By simply turning on the Websays crawler and automated sentiment analysis and get a popularity ranking very close to the real votes.