Overview of the We Feel Project

We Feel is a project that explores whether social media – specifically Twitter – can provide an accurate, real-time signal of the worlds emotional state.

It is a collaboration led by mental health researchers at the Black Dog Institute, computer scientists at CSIRO and information technology specialists Amazon AWS and GNIP.

Emotional tweets are captured, analysed, and displayed visually in real time on a website (wefeel.csiro.au). Data on the tweets and their emotional content is made available to the public for their direct use.

The online tool passively analyses a ‘massive pipe’ of Tweets up to 32,000 tweets per minute, or around 27 million tweets per day. We Feel uses language-processing techniques to look at the English words people use in these posts and then maps these words to a hierarchy or ‘wheel of emotions’.

Social media monitored

Twitter

1
1

"The power of this information cannot be underestimated."

− Professor Helen Christensen, Executive Director, Black Dog Institute

The Challenge

Each year, approximately one in every five Australians will experience a mental illness and almost half of all Australians will experience a mental illness in their lifetime. People talk about a lot of things on Twitter but mostly people talk about themselves. Around 6,000 tweets go out every second which is a lot of information that Australian researchers hope will help us understand how people across the world are feeling in order to better understand our mental health.

The Solution

‘We Feel’ is an online web tool developed by CSIRO for the Black Dog Institute together with the support of Amazon Web Services and GNIP, that aims to verify whether social media can accurately map our emotions. The ‘We Feel’ tool passively analyses a ‘massive pipe’ of Tweets up to 32,000 tweets per minute on average, or around 27 million tweets per day. It uses language-processing techniques to look at the words people use in their posts mapping these words to an emotion hierarchy or ‘wheel of emotions’, offering a summary view of the prevalence of different emotions across the world. In order to create ‘We Feel’, CSIRO researchers had to create an emotional map. To do this they set up a crowd-sourcing task on the internet and asked people to look at a sample of 600 words in our vocabulary that relate to emotions and asked them 1) is the word an emotion and 2) which emotions does the word best correspond to. The researchers then plotted these emotions across a four quadrant scale and selected colours that corresponded to the emotions (Yellow = Surprise, Green = Joy, Red = Anger, Blue = sadness for example) in order to design the user interface. We Feel allows you to explore each of these emotions and its sub emotions on a segmented coloured wheel and look at its prevalence visually as a river stream across a minute by minute time scale which extends back days or several weeks. Users can also explore emotions across locations around the globe and select other search criteria such as gender to further refine the results. The aim of this project is to validate the use of Twitter as a tool for observing real-time trends in emotional changes in our community. The tool will also help understand questions such as how strongly our 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. One of the key innovations in this tool is the ability for researchers to crunch such a large volume of Twitter data in real time. Amazon Web Services had provided CSIRO access to its latest big data analytics service, Kinesis, which allows for the real-time processing of streaming Twitter data coming down the pipe.

The Impact

Having access to real-time data is also of enormous benefit to mental health researchers. Currently, research and potentially life-saving public health programs are based on statistics that may be five years old and are based on snapshots of a particular place and time. Twitter offers a large and fast sample of information that could hold the key a real-time view of our emotions. This new information from We Feel will be compared to existing literature to see how social media can be used to contribute to the real-time tracking of mental health in Australia. This project has provided a prototype tool to support Black Dog Institute researchers to gather evidence on role of social media in the development of more effective public health campaigns as well as the evaluation of existing campaigns. 

See some of the Real-Time We Feel data

Problem we set out to Address

Tens of millions of tweets are posted every day by a variety of different people and on different topics. The people who tweet; what they are up to, what they have encountered, and how they feel about it is of great interest, with considerable potential for providing greater understanding of how emotions are experienced in real time.We Feel aims to clarify these shared thoughts via Twitter to better understand the prevalence and drivers of emotions both within Australia and internationally. Whilst there is already a wealth of academic research on mental health and wellbeing, such information is traditionally gathered by surveys and is unable to give a real-time indication of what is occurring on a day-to-day basis. This traditional approach is also time consuming and expensive. Via Twitter, We Feel offers a large and fast sample of information that could hold the key to a real-time view of our emotions.

Potential Uses

The We Feel project has the potential to identify locations with higher levels of depression and suicide risk, and fluctuations in risk over time and between genders. It could also help understand questions such as how strongly our 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. Harvesting tweets is a relatively new field, with huge potential, but at this stage our work needs to explore its possibilities and its limitations.

Contributors

Mark Larsen
Tjeerd Boonstra
Philip Batterham

Bridianne O’Dea
Cecile Paris
Helen Christensen
David Milne

Funding
  • Education Grant from Amazon Web Services
  • CSIRO
  • NHMRC John Cade Fellowship APP1056964