Researchers at Monash University have trained a computer to analyse 2.5 million tweets, providing insight into what people were most worried about during the coronavirus pandemic.
As public health measures involving strict lockdowns were introduced in countries around the world in March, people took to Twitter to have their say on them.
Australia tweeted about panic buying more than any other country, especially about toilet paper and limits on alcohol purchases.
RELATED: Word that stumped Insiders host
We also had a lot to say about decisions being made to restrict the number of mourners at funerals while hair salons remained open.
Along with the Irish, Australians were quick to label fines for breaching ambiguous new rules as “revenue raising”.
A team led by Monash machine learning researcher Caitlin Doogan created “topic models” to analyse tweets related to the implementation of public health measures in response to COVID-19.
“Topic models are algorithms that identify co-occurring words in texts, then group those texts together into collections that represent a ‘topic’,” Ms Doogan explained.
The team used a model that “specialises in accurately modelling topics on content found specifically in tweets”.
The analysis has yielded some surprising insights about the ways different countries reacted to non-pharmaceutical interventions designed to stop the spread of coronavirus.
New Zealand had some of the harshest lockdowns in the world, but online its citizens “showed broad community support and intent to comply”.
Meanwhile users in the US and Canada, where lockdowns were relatively mild, there was a “protracted debate over such restrictions and support for adherence was not as easily interpreted”.
The researchers concluded this was because the New Zealand government did a better job communicating, finding that lower levels of commitment to following restrictions “appear to be rooted in both the complexity of the imposed regimen and the corresponding lack of understanding of the regimen in the community”.
The study’s findings have been published online and accepted for publication by the Journal of Medical Internet Research,and while it’s yet to be peer-reviewed, the researchers said it “supports the hypothesis that public understanding and responsiveness is supported by consistency, clarity and timeliness of government messaging”.
The US tweets were the most “racially charged”, with almost all the posts containing “anti-China” hashtags coming from there.