Rain water level information with flood warning system using flat clustering predictive technique

Rain and typhoons are normal in the Philippines due to its climate, however, long period and intensified rainfall causes flood in some areas in the Philippines, especially in Metro Manila. This paper aimed to provide prediction on the status of flood level in the monitored streets with the use of a predictive model. The flood levels of designated streets are shown as well as the rainfall measurement. The microcontroller gathers all the data from the rain gauges and the online web server will provide a convenient user interface to be browsed. This server has also the system of warnings and advisories and has a flood projection page from which a person can tell the expected flood level based from a given rainfall measurement and rainfall duration. The predictive model provides its prediction based solely on the data being sent to it which are the rainfall measurement, rainfall duration, and past data of flood level. The system used Raspberry Pi, rain gauges, an online web server for the data presentations and a predictive model using Python programming language to project possible flood scenarios. Results of the study showed that the Raspberry Pi can support all the required hardware and software requirements for the system which makes it efficient and feasible. Color coded rainfall intensity and flood levels which follow the standards used by PAGASA deemed to be useful for a more user-friendly interface. The system has been tested to provide necessary information to the public for their preparation to incoming storms and typhoons.