How happy are you? New Twitter tool takes the pulse of the world's mood
Around a third of Tweets by Hongkongers made in the last 24 hours expressed happiness, according to a new online tool that maps moods around the world in real-time.
Australian researchers have unveiled “We Feel”, a Twitter tool which analyses up to 32,000 tweets per minute - about 10 per cent of all English-language tweets - for 600 words that are then linked to emotions such as love, joy, surprise, anger, sadness and fear.
The researchers explained that the app uses the 600-word vocabulary of “emotion terms” compiled from multiple sources, which were then organised through crowdsourcing into a “hierarchy of emotions” comprising “love”, “joy”, “surprise”, “anger”, “sadness” and “fear”. Public tweets mentioning one of these 600 words are mapped into one of the six terms and categorised by time and country or time zone.
For example, by 3pm on Tuesday, there were 64,005 emotional Tweets from Hong Kong within the last 24 hours, corresponding to about 59 per cent of all Tweets from the area in that timeframe (109,100).
The majority of the emotional Tweets, 36,142, were “joy”. This corresponded to less than half of the emotive Tweets and a third of all Tweets made in that time frame in Hong Kong.
Of the remaining Tweets, 8,550 were “sadness”, 4,254 were “love”, 1,957 were “surprise”, 1,826 were “fear”, 1,821 were “anger” and 9,455 were “other”.
In comparison, of Beijing’s 517,203 emotional Tweets in the same period, 282,483 were “joy”, 77,440 were “sadness”, 32,507 were “love”, 17,019 were “anger”, 15,912 were “surprise”, 15,089 were “fear” and the remaining 76,753 were “other”.
The data will be used to monitor the emotions of individuals and communities across different locations, and “ultimately predict when and where potentially life-saving services are required”, said lead researcher Helen Christensen of Australia’s Black Dog Institute, which researches and treats mood disorders such as depression.
“The power of this information cannot be underestimated. Currently, mental health researchers and associated public health programmes use population data that can be over five years old,” the professor and director of the institute added.
Researchers said they hoped it could uncover, for example, where people are most at risk of depression and how the mood and emotions of an area or region fluctuate over time.
Having access to real-time data could also help further understanding of how strongly emotions depend on social, economic and environmental factors such as the weather, time of day, day of the week, news of a major disaster or a downturn in the economy, they added.
They plan to submit a report to the New South Wales mental health commission at the end of the roject
The large volume of data from Twitter - which says it has 255 million monthly active users worldwide - is analysed with the support of Australia’s peak science body, the CSIRO, and internet giant Amazon’s remote computing services.
The project currently only trawls through English-language tweets, is limited to Twitter and is hindered by the lack of data about the gender, identity and location of some users on the social media platform.
The team said they used time zones to roughly locate accounts, focusing on Australia. However they said this method did not work as well in other places, for example North America.
Despite these limitations, project researcher Bridianne O’Dea said the tool was the first opportunity to get a greater understanding of people who use the platform to tweet about their emotions.
“This demonstrates that we can monitor people over time, we can pick up trends, and now it’s about validating this and see if these trends are indicative of what’s really going on,” said O’Dea, a post-doctoral research fellow at the Black Dog Institute.
“Now that we can collect data over time, we can do time comparisons and pretty much get a greater understanding of how people are using these technologies to express how they feel - because we don’t know that yet. Nobody knows that yet.”
The “We Feel” tool can be accessed at wefeel.csiro.au.