A new study from UC Irvine suggests that public health officials could use everyday conversations on social media to better understand how communities are affected by extreme heat and to develop more targeted interventions.
The research, published in Environmental Research Communications, analyzed 24,598 heat-related posts on Twitter (now X) from across California between 2016 and 2022. The study found that a person's living situation, transportation methods, and available resources were associated with how they discussed extreme heat, including its effects on their daily lives or their coping strategies.
The study was conducted by Nicolas Rodriguez Garcia and senior author Suellen Hopfer of the Joe C. Wen School of Population & Public Health, Veronica J. Berrocal of the Department of Statistics, and Chen Li of the Department of Computer Science. Hopfer, an associate professor of health, society and behavior, noted that "Social media gives us another window into how people are experiencing heat in their daily lives."
One significant finding emerged regarding housing. Counties with higher numbers of households living in RVs, vans, or boats—termed "sensitive housing"—were 15 percent less likely to post positive heat tweets and had 22 percent lower odds of posting about coping strategies. Researchers attribute this to potentially poor insulation in such homes, making them dangerously hot. This pattern suggests a need for targeted communication on indoor heat safety for these residents.
Transportation patterns also revealed vulnerabilities. Counties with more active commuters—people who walk, bike, or use public transportation—showed a greater likelihood of posting negative heat-related content and descriptions of heat's effects on daily life. Researchers note that these commuters face direct heat exposure, highlighting the importance of heat-conscious transportation planning.
Disparities were also identified in counties with larger Hispanic populations, which showed a slightly greater likelihood of posts describing heat’s impacts and a lower likelihood of posts on coping strategies, indicating an opportunity for more targeted awareness efforts.
The UC Irvine team employed a large language model to group thousands of posts and validated its accuracy against manually sorted posts, finding accuracy ranging from 74 to 90 percent. As heat-related posting increased with average monthly temperatures, the researchers envision social media as a complementary tool to traditional public health surveillance. It could offer immediate insights into heat exposure that are difficult to capture through conventional data alone.
Combining these social media signals with data on housing, transportation, and socioeconomic conditions could help public health agencies identify where specific heat communication is most needed. For example, communities with sensitive housing might benefit from messaging focused on keeping indoor spaces safe, while areas with many active commuters could benefit from interventions that reduce exposure during travel. This approach could lead to more tailored public health strategies as California prepares for more frequent and severe heat events.
