Analysis of Social Media for Social Good

Social Media Analytics for Social Good

Attitudes and Perceptions towards COVID-19 Public Health Measures

Led by Caitlin Doogan, this project seeks to address the need for insights about adherence to Public Health Measures (PHMs) during the COVID-19 global pandemic. Without a vaccine, the spread of COVID-19 can only be stopped with PHMs such as lockdowns, social distancing and face masks. But as fatigue sets in and compliance decreases, governments need new ways of encouraging people to adhere to PHMs. To do this, they need to know the public’s attitudes and understandings of these measures are and why. In this research, we used machine learning methods to analyse millions of tweets from six different countries. We wanted to know what PHMs people were talking about and what factors would potentially reduce or increase their adherence to them. This information will help governments make informed decisions about how to implement these measures and ultimately, stop the spread of COVID-19. We have developed an interactive website to explore the related data from this project.

The project team is Caitlin Doogan, Wray Buntine, Henry Linger and Samantha Brunt.

Media coverage

Taylor, J. (Aug 28, 2020). Twitter study shows Australians focused on panic buying as US users blamed China for COVID-19. The Guardian. Available here.

Tecregister. (Aug 28, 2020). Twitter study shows Australians focused on panic buying as US users blamed China for Covid-19 | Australia news. Techregister. Available here.

Griffith, C. (Aug 28, 2020). Coronavirus: Twitterati trends air world of confusion. The Australian. Available here.

Australian Associated Press. (Aug 28, 2020). Tweets reveal what we think of COVID rules. 7news. Available here.

Watch Our Overview of the Project

Research Story: How We Explored Millions of Tweets

Click on the picture to view

COVID-19-NPI-visualisation.png

What Australia Tweets about #COVID19


Analysing Change for a Lower Carbon Future

Led by Professor Wray Buntine, this project is part of the Woodside Monash Energy Partnership, and is a collaboration between the faculties of Law, Business and Economics, and IT, as well as the Monash Sustainable Development Institute (MSDI). The FIT team is Wray Buntine, Caitlin Doogan and Richard Le. The project employs advanced deep learning and qualitative methods to mine insights from our complex media ecosystem. This research aims to inform organisations such as Woodside Energy, about public knowledge and sentiment toward hydrogen transitions in Australia and how best to move toward a low-carbon future.

Analysing Change for a Lower Carbon Future

Ten years of the ‘Right to be Forgotten’ on Twitter

Led by Caitlin Doogan, this project develops hybrid computational methodologies to explore ten years of Twitter discourse about the ‘Right to be Forgotten’ (RtbF), which is now part of the GDPR. News is increasingly sourced from social media, making it essential to determine how the traditional, digital and social media influences public understanding and discussion around matters of policy and regulation. This work offers a new pathway for the empirical analysis of social media to gain an understanding of the interconnectivity in discussions like these. We demonstrate how the granularity of computational analysis combined with intertextual interpretations highlight how new meaning emerges from the complex dynamics between online and offline media systems.

The project team is Caitlin Doogan, Henry Linger, Wray Buntine and Annelie de Villiers.

Ten years of the ‘Right to be Forgotten’ on Twitter

Designing effective classifiers for improved online tracking of campaign promises in the 2019 Indian general election and 2020 US Presidential elections.

This interdisciplinary project is led by Prof. Robert Thomson and is part of a broader collaboration between Monash University and the University of California San Diego. It examines processes that lie are at the heart of democratic practice through the assessment and application of deep learning techniques and data mining practices. For democracies to function effectively, political candidates must offer voters meaningful choices, in the form of campaign promises, during election campaigns. Given the increasing use of social media in elections, we need new ways to mine posts about these promises from social media platforms, which cannot be achieved by traditional means. This project aims to develop new methods with which to achieve this. The FIT team is Wray Buntine,  Caitlin Doogan  , Panatchakorn Ananthothai, and Pankaj Adhikari.

US Presidential elections 2020