Semantic analysis on microblog data for emergency response in typhoon Chan-hom

Abstract

Social media play an important role in emergency response. With Twitter data were widely used in emergency management, this paper attempted to extract valuable information from microblog data to support emergency response. Event-related microblog data were collected from the Sina Weibo website. Data mining and semantic analysis methods were applied to obtain citizens' risk perception and public opinion from the collected data. Using Typhoon Chan-hom, this paper considered public opinion from the spatial and temporal perspectives. The citizens' risk perception was shown based on the geolocation information. Semantic analysis was conducted to analyze the change of public opinion in three provinces. Two stages were divided based on the posting time during crisis. By comparing the most common keywords at different stages, public opinion and people's requirements were extracted to support decision-making. This paper provides a new insight to emergency response in China.