Emotive, a new computer software designed by British scientists at Loughborough University, aims to map the mood of the nation by analysing emotions of Twitter users.
It is claimed that the program can scan and analyse up to 2000 tweets per second and extract a particular emotion from each one. The eight basic emotions it claims to detect are: anger, fear, disgust, shame, surprise, happiness, sadness and confusion. Apparently Emotive can geographically assess the mass mood of the nation and could help the police to prevent potential criminal behaviour or threats to public safety.
This seems like good news after Twitter was criticised for apparently being a key player in assisting in organising riots that shook the nation in the summer of 2011. Riots were organised partly through collective tweeting on the website, and had this program been available then, those threats and clear warning signs could have been detected, possibly preventing such large groups meeting or moving riot police to key areas so they could act fast.
Although a computer program won’t be able to be specific or especially accurate, it could be a great tool when analysing a large amount of data, such as a general uproar amongst Twitter users or trending of certain topics that are bringing waves of emotions.
However, unlike a human being, the program can’t focus on individual tweets to correctly identify the user’s emotions, which may lead to confusion and possibly flawed results. For example, can it detect sarcasm? Jokes? Or fake Twitter accounts that post automated tweets? I think we can all agree that sarcasm, satire and strange fake accounts are what makes Twitter. That and the trolls, of course. Therefore genuine data may get mixed up with sarcastic Twitterers, proclaiming their utter delight at the rain outside, those excellent football results or that fantastic boyband song at number one.
With more than 500 million people using Twitter and 340 million tweets posted every day, there’s bound to be a flaw in such a huge undertaking. However, for now the researchers insist that only tweets in the UK will be analysed, although plans to go worldwide are in the works. So maybe it won’t be able to correctly identify happiness relating to popular culture, as sarcasm runs deep through Twitter’s veins, however regional or national upset that is projected through tweets may be able to be identified. In the long run, analysing the emotions largely shared on Twitter over news or politics, then working with local authorities to tackle issues head on, could make a real difference in ensuring that our tweeting nation is a happy one.
