A high degree of real-time interconnectedness can aid information transmission, particularly in disaster situations. However, it can have substantial negative consequences when information is emotionally laden and transmits these emotions, particularly the emotion of panic, to the individual across social media in an already grave situation. Prior research has shown that information laden with emotion spreads through social network faster than otherwise. Hence, we highlight the need to understand and curtail potentially panic-causing information, without compromising on good quality information from being available for effective crisis communication and management. With this research, we present the necessity of detecting the panic potential of social media messages, and aim to address two research questions: What are the features, and metrics necessary, to compute and evaluate the panic potential of a social media message (respectively)? Our planned analysis takes the case of the Munich shooting incident, 2016, based on user tweets immediately after the incident. Different features and evaluation metrics are proposed and discussed. The work aims to detect panic potential of messages in social media networks during disasters.

First Page


Last Page