Social media services like Twitter and Facebook don’t just have to be used for communication. They can also be used to predict epidemics.
The next time you post to Facebook or Twitter, or any other social media service, you might want to think about how you use your words.
While these words represent you, they may do a little more than that in the coming years. They might be used as an indication of what’s to come, as researchers tap into the data science behind social media.
It’s something that has come out of Australia’s government science organisation, the CSIRO, with its data science arm, Data61, developing a tool to scour social media and work out events ahead of time.
Developed by Data61 in collaboration with UNSW’s Kirby Institute, the team has built a tool that uses artificial intelligence to scan for tweets that match specific words and phrases in locations after an event, and then process the informational with natural language processing to highlight situations that can be grouped together. The tool processes the words with AI and works with time modelling to work out which social media posts might be referring to a specific event.
For the test, the tool was used as way to detect an asthma epidemic that it Melbourne three years ago. The asthma attack in question saw over 8000 hospital admissions on November 21, 2016 following a local thunderstorm, with a type of asthma known as “thunderstorm asthma”. While the technique being used to track this specific condition is being used in a post-dated way, because posts made to Twitter are archived typically so that everyone can see them (unless user privacy settings are set to high), the team from Data61 and UNSW Sydney was able to analyse over three million tweets using words like “breath” and “coughing” related to asthma to test its tool.
Using this process, the team was able to prove its system could have predicted the medical event ahead of time, and could be used as a forecasting tool to help hospitals and organisations be as prepared as possible.
“In future, this system can be used to provide health authorities and the community early warning of a serious and sudden health event,” said Professor Raina MacIntyre, Head of the Biosecurity Research Program for UNSW Sydney’s Kirby Institute.
“Early detection could significantly improve our capability to mitigate the impact of epidemics.”
While it’s not yet known if Australian organisations will use or embrace the method, it adds to Australia’s growing contributions to the world of science and technology, and highlights the great ideas the country has added to the world.
