A Carleton University researcher is using artificial intelligence to mine local residents’ social media posts in an effort to identify people who lack food or medication during the COVID-19 crisis or might be showing signs of the illness but haven’t been checked.
Mohamed Ibnkahla is a computer engineering professor at Carleton and the university’s senior Cisco Industrial Research Chair in Sensor Networks for the Internet of Things. He and his team of researchers that includes about 20 students are using big-data analytics to conduct “sentiment analysis” of platforms such as Twitter and Facebook – that is, using AI to analyze social media posts and find out what concerns people most about the novel coronavirus.
Ibnkahla says the system can help “detect what is happening in real time” based on what people are saying on their social media channels, such as whether they’re running low on medicine or food or whether they might be showing symptoms of COVID-19 but haven’t been tested.
A statement from the university says the process could be used to help predict when and where shortages of food or medicine might occur as well as “identify false rumours that are gaining traction on social media, so that public health officials can address them directly.”
Ibnkahla says he believes the data could give public health officials a more complete picture of the true scope of COVID-19 in the community, allowing them to better project the coronavirus’s path.
“I expect more accurate predictions and probably closer to reality because we have much more information,” he says.
Ibnkahla says residents’ levels of anxiety and pessimism regarding the virus have ebbed and flowed somewhat over the last couple of months.
In February, he says, social media users tended to be very negative and anxious as they worried about potential shortages of items such as food, medicine and masks. Toward the end of March, people began to adopt a slightly more optimistic outlook as the government started to unveil programs designed to aid struggling businesses and laid-off workers.
“These are things that we can learn from the data,” he says.
Ibnkahla, who is currently conducting his study on a volunteer basis, wants to build a system that can be updated daily to help government officials address the public’s areas of greatest concern about the virus. He hopes the program can be expanded to other parts of the country, but says that would require additional funding.