Exploring Public Reaction to Air Quality on Social Media: The 2022 Saharan Dust Episode in Lisbon

Vollständiger Abstract

Worum geht es in dieser Arbeit?

This study analyses the relationship between air quality and sentiments expressed on social media, focusing on the Saharan dust episode that affected Lisbon in March 2022. Environmental data were collected alongside geolocated Twitter posts (now rebranded as X) from the Lisbon area. The content of these posts was analysed using the Valence Aware Dictionary and sEntiment Reasoner algorithm to identify emotional patterns and assess correlations with fluctuations in atmospheric pollutants. A predictive analysis was also conducted using Decision Tree and Random Forest machine learning algorithms. The findings suggest that despite high levels of inhalable particulate matter (PM10) during the event, there was no significant change in online sentiments, which remained mostly neutral or slightly positive. This implies that the publics environ-mental perception may not be clearly reflected on social media, possibly due to cultural factors or the limited influence of Twitter in Portugal. The study concludes that integrating environmental and social data is feasible and valuable, but further research is needed to explore longer periods, other cities, and more advanced techniques, thus contributing to the understanding of air pollution and public perception in Portugal.