Predicting Moods Using Social Media
Ever worry what people think of your brand? Or the general mood of your neighbours and district? What if there was a way to find out? Well, it's on our doorstep!
Some pretty fancy-pantsy scientists at the University of Bristol have come up with a nifty way of keeping track of peoples emotions. Actually it's less tracking, more about getting a general feel.
Thomas Lansdall-Welfare and co. analysed nearly 500 million tweets from 2009 to 2012, arranging them by joy, fear, anger and sadness.
Soon they were able to create a complicated looking graph and an incredibly moving experience, as science often does.
By pairing the graph with something called the 'Grimace Project', they were able to create a very personable face that reflected the data of the UK tweets (see above.. not below).
You watch his face wiggle up and down (he has particularly excited eye-brows) with the general mood of the UK from 2009 to 2012.
It's warming and relatable to see a human face representing the entire mood of the UK, especially when he starts smiling around Christmas.
Equally, it's terrifying to see the face of this helpless individual terrified by the messages uploaded to the cloud around the time of the Riots and Budget Cuts.
[S]ome reduction in fear and anger in the run up to the royal wedding in April 2011, suggesting Prin[c]e William and Kate Middleton's impending nuptials had a calming effect on the public mood. Following the wedding there was a sharp rise in anger, culminating at a peak on 6 August 2011, the date of the first UK riot. - Jacob Aron, New Scientist
Firstly then, we may be able to use this sort of data analysis to gauge the general opinion on certain subjects, topics and brands. More interestingly though we can use it to pick up on trends, predicting the general mood of the people, use it as an early warning system.
I see great potential in these technologies and meta-analysis methods. But with most new technologies in the Social Media sphere we need to take it with a pinch of salt and approach it with caution.
This sort of meta-analysis isn't going to be widespread for a while yet and we have to remain vigilant that such mass surveillance isn't wielded immorally.
Still, very exciting stuff.
Other Posts by Matthew Harris
Social Media Today