IN23C-1743
Big Data Analytics for Disaster Preparedness and Response of Mobile Communication Infrastructure during Natural Hazards

Tuesday, 15 December 2015
Poster Hall (Moscone South)
Lei Zhong1, Kiyoshi Takano2, Yusheng Ji3 and Shigeki Yamada3, (1)Earthquake Research Institute, University of Tokyo, Tokyo, Japan, (2)University of Tokyo, Bunkyo-ku, Japan, (3)National Institute of Informatics, Tokyo, Japan
Abstract:
The disruption of telecommunications is one of the most critical disasters during natural hazards. As the rapid expanding of mobile communications, the mobile communication infrastructure plays a very fundamental role in the disaster response and recovery activities. For this reason, its disruption will lead to loss of life and property, due to information delays and errors. Therefore, disaster preparedness and response of mobile communication infrastructure itself is quite important.

In many cases of experienced disasters, the disruption of mobile communication networks is usually caused by the network congestion and afterward long-term power outage. In order to reduce this disruption, the knowledge of communication demands during disasters is necessary. And big data analytics will provide a very promising way to predict the communication demands by analyzing the big amount of operational data of mobile users in a large-scale mobile network.

Under the US-Japan collaborative project on ‘Big Data and Disaster Research (BDD)’ supported by the Japan Science and Technology Agency (JST) and National Science Foundation (NSF), we are going to investigate the application of big data techniques in the disaster preparedness and response of mobile communication infrastructure. Specifically, in this research, we have considered to exploit the big amount of operational information of mobile users for predicting the communications needs in different time and locations. By incorporating with other data such as shake distribution of an estimated major earthquake and the power outage map, we are able to provide the prediction information of stranded people who are difficult to confirm safety or ask for help due to network disruption. In addition, this result could further facilitate the network operators to assess the vulnerability of their infrastructure and make suitable decision for the disaster preparedness and response.

In this presentation, we are going to introduce the results we obtained based on the big data analytics of mobile user statistical information and discuss the implications of these results.