Sensor data collection and analytics with thingsboard and spark streaming

The diffusion of small low cost sensors has opened new opportunities for the design of real-time monitoring systems in several application fields. Internet of Things (IoT) is a new branch of Information Technology that connects several of these heterogeneous devices through different network protocols to provide large scale interoperability. In this paper we describe how some open source software tools can be integrated to collect, monitor and process streams of data received in real-time by sensor devices. The tools we have employed are the Things-Board platform to collect sensor data and the Spark Streaming framework for cluster computing to perform data analytics. The presented architecture can be exploited in various monitoring scenarios (such as clinical, environmental or energetic processes) to study the trend of some key parameters and also to produce alerts in case of problematic situations. Without loss of generality, we focus on biomedical data about the breath of patients affected by chronic respiratory disease: real-time monitoring of breath parameters enables automatic ventilotherapy and allows also to warn doctors in time in severe cases.