368 / 2019-02-07 08:09:33

An analysis of Factors Influencing Disaster Mobility using Location Data From smartphones: Case study of Western Japan Flooding

flooding,mobility,smartphone

全体主题 > Big data in Urban Planning and Management

Western Japan was hit by heavy rain from 28th July to 8th of June, 2018. Record-breaking rain caused rivers to burst their banks in Hiroshima and other areas. Over 200 people were died because of this. After heavy rain, the authorities tried to grasp why evacuation would be late. They men-tioned that normalcy bias and cognitive dissonance are one of cause of a lot of damage [16]. Also, effective alert system is demanded to make people do evacuation behavior at an appropriate time. To understand which factors have influence to people’s behavior, we estimate the probability of irregular behavior by one unit change in external condition. We choose 500m mesh as a unit of analysis to take into account individual singulari-ty and classify types of mesh as 3-classes to figure out abnormal behav-ior. We verify that the more residents each mesh has, person in that region is more likely to do the action of normalcy bias. Because of limitation of data, accuracy of this method is a bit low. However, there’s a change that administrative officials could choose adequate alert system. Using prediction’s result of people’s mobility and information about risk of disaster, they could deal with this dangerous situation well.