A big-data-based Urban flood defense decision support system

As cities in developing countries are expanding rapidly in recent years, flood has an increasing impact on urban management. In this paper, we present the design and implementation of an urban flood defense decision support system based on big data. The system connects real-time sensor to collect streaming data, and uses a data-driven method that considers temporal and spatial factors to forecast water level in the next 6 hours. Thus, it can provide enough time for the authorities to take pertinent flood protection measures such as evacuation. Our predictive model is a hybrid of linear regression and artificial neural network, and can give early warning of potential flood using the forecast results. The system is implemented on Java EE platform, and integrated with Baidu Maps API to provide a user-friendly interface. © 2015 SERSC.