Floods impact dynamics quantified from big data sources (1804.09129v1)

Abstract: Natural disasters affect hundreds of millions of people worldwide every year. Early warning, humanitarian response and recovery mechanisms can be improved by using big data sources. Measuring the different dimensions of the impact of natural disasters is critical for designing policies and building up resilience. Detailed quantification of the movement and behaviours of affected populations requires the use of high granularity data that entails privacy risks. Leveraging all this data is costly and has to be done ensuring privacy and security of large amounts of data. Proxies based on social media and data aggregates would streamline this process by providing evidences and narrowing requirements. We propose a framework that integrates environmental data, social media, remote sensing, digital topography and mobile phone data to understand different types of floods and how data can provide insights useful for managing humanitarian action and recovery plans. Thus, data is dynamically requested upon data-based indicators forming a multi-granularity and multi-access data pipeline. We present a composed study of three cases to show potential variability in the natures of floodings,as well as the impact and applicability of data sources. Critical heterogeneity of the available data in the different cases has to be addressed in order to design systematic approaches based on data. The proposed framework establishes the foundation to relate the physical and socio-economical impacts of floods.