A Study on the Monitoring of Complex Disaster Using Crowd Source Data

As industrialization and urbanization progress around the world, more complex and large-scale complex disasters are occurring, causing numerous casualties and property damage every year. As climate change gradually accelerates and its impact grows, such as recent cold waves and heavy snow in the United States and abnormal temperatures in Europe, it is difficult to predict with existing physical modeling alone. Recently, disasters are gradually expanding in the form of covering not only natural disasters but also various social disasters. Social disasters cover disasters such as fires, infectious diseases, and fine dust caused by human activities. Unlike natural disasters, it is difficult to measure numerical values and predict occurrence patterns in real time, so it is very important to respond quickly through information sharing. There is a limit to establishing the same response system globally to respond to disasters that may occur worldwide, so it is necessary to develop a platform that can quickly share cases while being economical. With the recent development of communication technology, about 70% of the world's population uses smartphones, and various unstructured data are being generated in real time through various social media channels. Individuals act as a sensor and can share their location or current situation in real time. Therefore, the purpose of this study is to develop crowd sourcing technology using social media, analyze the collected data, and present ways to use it in the event of a disaster. In this study, a platform was established to collect and analyze disaster-related SNS data such as floods, fine dust, and forest fires, and it was designed so that users could receive information through websites and apps. As a result of application to various disaster cases in Korea, the temporal and spatial correlation between disaster occurrence patterns and social media data was high, and the possibility of using initial monitoring methods was proved. This result can be applied to all disaster disasters or crimes, and it is expected to be highly useful as it can quickly verify disaster thoughts and share cases in real time.