Twitter mood tool gauges post-Budget feelings

A new CSIRO project using big data analysis tools from Amazon Web Services aims at tracking the mood of the English-speaking world.

Australians were significantly more emotional after Tuesday's Budget announcements, according to a new research tool released by the CSIRO, in conjunction with the Black Dog Institute and Amazon Web Services.

The We Feel tool developed by CSIRO flags approximately 600 specific words and assigns tweets to certain emotions such as fear, anger, surprise, and sadness. It was funded by CSIRO, the National Health and Medical Research Council, and Amazon Web Services.

The online tool, built on Amazon Web Services' Kinesis big data analysis platform, analyses around 32,000 tweets per minute across the English-speaking world using a random 1 percent sample from Twitter's public API, a random 10 percent from Gnip, and a third source that monitors the Twitter public API for a vocabulary of what CSIRO believes to be emotional terms.

AWS' EC2 captures the tweets and annotates with emotion, gender, and rough location that is used to build up summaries required for the graphs produced on the website. The site itself was made using AngularJS and Bootstrap, with the visualisations constructed using D3.

The tweets on Budget day compared to the Tuesday before Budget. Image: Screenshot by Josh Taylor/ZDNet

According to the We Feel tool, there were approximately 68,000 more Tweets in the Oceania region on Budget day last Tuesday compared to the previous Tuesday. The number of emotional tweets peaked at 8pm on both days, and despite the strong reaction to the Budget on Twitter, the CSIRO tool recorded largely the same emotional response on both days.

The project's lead researcher Helen Christensen said that the ultimate goal of the We Feel tool is to help provide mental health organisations with the ability to intervene when people may be at risk of suicide, or self harm.

"The end game is to be able to respond in real time to people in high-stress situations," she said.

This may come in the form of a tweet from an organisation such as Lifeline to a Twitter user who is tweeting emotively.

The project still has a long way to go before it reaches that stage, however. Researchers admitted that at this point there was no way to tell whether a person is being sarcastic or flippant in the emotional word used in their tweet.

The data collected by the CSIRO is shared, and can be used for non-commercial purposes for free.