Brits Turn to Twitter to Gauge National Mood


A team of Loughborough University researchers has developed software designed to answer one simple question: How are the British feeling today?

The program is called EMOTIVE -- an acronym for Extracting the Meaning of Terse Information in a Geo-Visualisation of Emotion. EMOTIVE is a linguistic sentiment analysis tool that scans UK-based Twitter posts at a rate of up to 2,000 Tweets per second. Using a specially developed ontology, EMOTIVE assigns scores to individual Tweets in eight categories, assessing levels of the following emotional states:

  • Anger
  • Confusion
  • Disgust
  • Fear
  • Happiness
  • Sadness
  • Shame
  • Surprise

In addition to different parts of speech and sentence, EMOTIVE recognizes and understands hashtags and emoticons -- going beyond basic linguistic sentiment analysis.

Of course, accurately detecting and analyzing a person's mood based on their social media posts is nothing new. Researchers have done it many times before, demonstrating that a social media user's content can be used to predict not just a wide variety of emotions, but also facts about the user's identity -- such as gender, location, political affiliation, and other demographic factors.

Linguistic sentiment analysis of Tweets has also been used to accurately predict all sorts of future behavior in other contexts, ranging from movie box office receipts to election outcomes. Most notably (and, perhaps, commonly), it has long been employed to successfully predict stock market behavior.

In a particularly compelling example, computer scientists discovered in 2010 that the level of "calm" detected in Tweets via linguistic sentiment analysis could predict the stock market with 87.6 percent accuracy as many as six days in advance. This research formed the basis of the trading strategy used by Derwent Capital Markets, a hedge fund that invested clients' money solely based on Tweets.

Although Derwent's fund was short-lived and did not yield the 15 percent to 20 percent returns that it boasted that it would, it was still deemed a success. The fund's reported 1.86 percent return outperformed both the overall market and the average hedge fund.

Despite its potentially broad range of applications, EMOTIVE, too, seems to have a specific purpose in mind. Researchers on the EMOTIVE project say that the software will be able to help law enforcement geographically track potential criminal activity and public safety threats. Additionally, they posit that the British government will be able to use EMOTIVE to make policy decisions on national security matters.

2011 British riots
2011 British riots

Indeed, the project is partly funded by the UK's Defence Science and Technology Laboratory, a Ministry of Defence agency. It is, presumably, what the British government hopes is an answer to its prayers since the 2011 England Riots. At the time, the government was beside itself over how social media played into the equation. Rioters had used Twitter and other social networks to incite, plan, and brag about looting and other illegal activities. (It's possible they even used an app to escape police crowd control tactics.)

Prime Minister David Cameron had even suggested shutting down social media in response to future demonstrations of civic unrest. With EMOTIVE, however, it appears the British government's goal is to work with social media to stay a step ahead of criminals.

It is unclear to what extent each of the eight emotional factors EMOTIVE measures will be helpful in predicting events. The project remains a work in progress. The EMOTIVE team's next step is a prototype that the will purportedly automatically detect events while gleaning even more information from Tweets.

Think EMOTIVE has potential? Share your thoughts below.

Related Links:

Joe Stanganelli, Attorney & Marketer

Joe Stanganelli is founder and principal of Beacon Hill Law, a Boston-based general practice law firm.  His expertise on legal topics has been sought for several major publications, including U.S. News and World Report and Personal Real Estate Investor Magazine. 

Joe is also a communications consultant.  He has been working with social media for many years -- even in the days of local BBSs (one of which he served as Co-System Operator for), well before the term "social media" was invented.

From 2003 to 2005, Joe ran Grandpa George Productions, a New England entertainment and media production company. He has also worked as a professional actor, director, and producer.  Additionally, Joe is a produced playwright.

When he's not lawyering, marketing, or social-media-ing, Joe writes scripts, songs, and stories.

He also finds time to lose at bridge a couple of times a month.

Follow Joe on Twitter: @JoeStanganelli

Also, check out his blog .

Clean Your Data with Visualization and Algorithmic Tests

Speakers at Bio-IT World explore techniques for biotech researchers and others working with big data to identify the accurate data in their data files.

Data Sharing: A Matter of Life and Death

Cooperation among medical researchers -- done right -- very simply can mean lives saved, but the research community needs education on how to execute on that collaboration.


Predictive Value?
  • 9/16/2013 9:44:42 AM
NO RATINGS

I'm a bit dubious when various experts claim to have come up with ways to predict otherwise unpredictable events or human behavior. Predicting the stock market, of course always gets one's attention. There have been tens of thousands of those claims. But, in reality it's easy to pick and choose what one's data seems to correlate with. But correlation doesn't neessarily mean it's predicting anything more than a random chance that the data seems to match up. A broken clock is correct twice a day.

Re: Predictive Value?
  • 9/16/2013 12:52:56 PM
NO RATINGS

Agreed. When two events occur in succession, it does not necessarily make the events causal. And is a broken clock correct twice a day? Hahaha, that got me thinking.

Speedy call to action
  • 9/16/2013 8:22:31 PM
NO RATINGS

Joe, it seems so much of the UK government's success in using EMOTIVE would be not only in the ability to predict behavior but to head off such behavior -- all in real time or near to it as possible. Would you say that's a fair assessment? And, if so, how realistic is it?

Re: Predictive Value?
  • 9/17/2013 9:30:37 AM
NO RATINGS

I agree I am also very skeptical about this and its real value. I just think we are not that easily predicted.

Re: Predictive Value?
  • 9/17/2013 9:53:37 AM
NO RATINGS

I'm wondering how accurate this could be as well.  I do think it could be used for specific issues and if someone wants to look at multiple issues and make a call on the mood of the nation I don't see that being much different than a phone survey for example.  I do see a value in looking at individual's social media activity though especially teens since behavior changes can signal bigger issues like the potential to commit suicide.

Re: Predictive Value?
  • 9/17/2013 11:41:25 AM
NO RATINGS

So much of what I see taking place on Twitter relates to marketing of brands, companies, events, and so on so it's difficult for me to think of Twitter as a real source of information about individuals themselves. I know it exists, and is worth playing around with from an analytics/sentiment perspective, but I'm not sure how seriously we can take any of those activities at this point. And will all this effort be for naught, anyways, when the next big social phenomenon comes into play and people stop tweeting so much? Hopefully not!

Re: Predictive Value?
  • 9/18/2013 7:28:21 AM
NO RATINGS

I think that if the systems and software are built correctly then it won't matter what the next social media toy is it will just be a matter of scraping data from that source instead of Twitter.  I think that behaviors will be more important to track on social media sites than the words used.  A flurry of activity near the launch of a new product might signify a sponsored account.  A consistent stream of activity that references multiple sources might be more likely to be a fan/hobbyist. 

Freedom of speech
  • 9/18/2013 7:47:38 AM
NO RATINGS

" Researchers on the EMOTIVE project say that the software will be able to help law enforcement geographically track potential criminal activity and public safety threats"----- That may violate people's freedom of speech. Also, I wonder how many criminals would talk about their "potential" criminal activities on Twitter before commiting them.

Re: Predictive Value?
  • 9/18/2013 11:57:43 AM
NO RATINGS

SaneIT -- I hadn't thought about that, but you make an excellent point. So companies need to be certain that the social media analytics platform they select will be easily upgradable to support whatever comes next, and can analyze not only by sentiment but behavior. 

Re: Freedom of speech
  • 9/18/2013 12:08:08 PM
NO RATINGS

Hospice, can you explain further? How might law enforcement's use of tweet analysis potentially violate freedom of speech? I'm not sure I follow.

Page 1 / 4   >   >>
INFORMATION RESOURCES
ANALYTICS IN ACTION
CARTERTOONS
VIEW ALL +
QUICK POLL
VIEW ALL +