Center for Embedded Network Sensing

Sensor networks generate a variety of data streams in different temporal and spatial resolutions. The data come as numbers text, images, and audio and are dynamically produced periodically and sporadically. How can we organize hundreds of thousands of such data streams? How can we make it easy for scientists and engineers to publish and share such data streams? In this paper, we present our solution, SensorBase.org. It is a web application that not only provides the user with the functionality of a traditional database management system, but also runs under the notion of a Web 2.0 data experience with a responsive user interface design and RSS data feed techniques. SensorBase.org also aims to be a data search engine to promote exploration. Like a web search engine, the user should be able to search for structures, or rather, “signals”, in the data using simple language queries. We provide a solution to a specific type of signal search problem and describe a search framework derived from the solution. A visualization component is in development to help people understand the data and to enhance the results provided by a data search engine. The development of SensorBase.org allows us, the statisticians, to influence “data practices”. People share data and add documentation within the guidelines of SensorBase.org which supports good experimental design and helps future modeling efforts. Last but not least, SensorBase.org provides a base for data integrity and data quality.