Geovisualization and correlation analysis between geotagged Twitter and JMA rainfall data: Case of heavy rain disaster in Hiroshima

Abstract

With the development of wireless communication technology along with emergence of location-enabled mobile devices and cyber-physical social sensor networks, nowadays we can collect and store a large amount of geospatial data such as weather phenomena, human mobility, and social networking activities along with time of occurrence. Information or knowledge extraction from this so-called Big Data is a challenge to many geospatial information users due to the nature of data complexity and large data volume. In this article, we discuss handling of Big Data with GIS by utilizing 10-minute intervals of Japan Metrological Agency (JMA) rainfall data synchronized with Twitter messages by specific keywords like “rain” or/and “landslide” to analyse the relationship between environmental phenomena and social responses in heavy rain conditions in the Hiroshima region. The results were analysed and visualized through a geovisualization technique to evaluate the possible use of inputs from social media websites to the government decision-making process, especially for disaster and emergency preparedness in the near future.