Could AI use social media posts to predict heat stroke?

Social media can be a noisy place, more often associated with online misinformation than with supporting public health. However, researchers in Japan have found a way to use machine learning technologies to detect signals in the noise.

In a new study published in Scientific Reports, scientists used deep learning models to identify heat stroke in real time, presenting opportunities for early intervention and health system preparedness.
Heat stroke poses a significant health risk, especially during periods of extreme temperatures. As global temperatures rise, and the frequency and severity of heatwaves increases, the incidence of heat stroke is expected to soar. However, the impact of this elevated risk can be better managed if service planners have advance warning of a spike in heat-related illnesses.

Professor Sumiko Anno from Sophia University, Japan, led a team that analysed social media posts and transformer-based learning models to detect heat stroke risks in Nagoya City, Japan. A ‘transformer’ is a type of neural network architecture that learns context and meaning by tracking relationships between words. Developed by researchers at Google almost a decade ago, it is well suited to analysing social media posts.

The team used this deep learning model to identify tweets containing the word “hot” in Japanese. The team successfully collected about 27,040 tweets over a five-year period. By preprocessing the text data and applying advanced deep and machine learning techniques, the models were trained and fine-tuned to identify tweets related to heat stroke events.

Through mapping the locations of heat stroke-related emergency medical evacuations and matching them with geo-tagged tweets, the study demonstrated how social media data could provide an early warning system for heat stroke risks in urban environments.

‘By leveraging social media posts, we can enhance public health surveillance systems and facilitate the early detection of heat stroke risks,’ Prof Anno said. ‘Our findings emphasize the importance of real-time data monitoring to combat the health challenges posed by climate change.’

This research opens the door to future applications of deep learning and social media posts for real-time monitoring. Looking ahead, the team plans to establish an early warning system for heat stroke in Aichi Prefecture, with the aim of eventually expanding this system to a nationwide alert system for Japan.
‘Our methodology can be extended and adapted for monitoring emerging and reemerging infectious diseases, broadening its application in public health surveillance,’ Prof Anno said.