In recent years, the development of Internet of Things (IoT) technology has been remarkable. To utilize IoT technology to prepare for disasters, it is important to link the natural science approach, which focuses on observation and prediction, with the social science approach, which focuses on information utilization and systems. The authors proposed the construction of the IoT‒Community Observation System (IoT‒COS) as a framework for studying the operation and utilization of local observations with the participation of local residents. In this study, local observation, which is expanding along with the improvement of IoT technology, was analyzed from the perspectives of natural science (risk assessment) and social science (social application) using six case studies on community disaster prevention in Japan. Although risk assessment depends on the type of disaster, social applications also differ from risk assessment. In addition, while the purpose of risk assessment is to assess risk, it is also important to build relationships among the parties involved. An even greater variety of local observation efforts has been identified worldwide. It is necessary to more carefully and actively discuss the social positioning of observation networks in next-generation disaster management.