User Reactions About Wildfires on Twitter

Abstract

Forests are the most important part of nature that provides the global balance within the ecosystem. Therefore, wildfires are one of the natural disasters that mostly affect the ecological balance. As an interdisciplinary study, the aim of this study is to measure the reactions of users by classifying comments about wildfires on Twitter with machine learning methods and to investigate the measures against wildfires. In the study, the user comments on wildfires were used on Twitter, which is used by all segments of the society and provides data analysis. A pre-processing has been firstly made for the comments about wildfires by performing word-based text analysis. Sentiment analysis has been realized as positive, negative, and neutral. Moreover, each sentiment group has been evaluated by dividing into four mostly expressed categories. The classification model accuracies have been compared by analyzing with the standard statistical scales. In the study, 58% of Twitter users wish that the wildfires would be ended immediately, approximately 34% of users think that firefighting related to government is enough, 7% of users think that the firefighting is insufficient. Moreover, all Twitter users have frequently referred to firefighting, global warming, support, and sabotage probability in their comments for wildfires. This research supported with sentiment analysis, reveals that wildfires create an alarming situation for all segments of society and it is necessary to act together against wildfires.