Poster Abstract: Topological Analysis for Knowledge Discovery from Building Sensor Data

Distributed sensor networks provide greater detail and potentially valuable insights into the behavior of complex systems, for instance retrofitting buildings with heating, ventilation and air conditioning schemes to improve efficiency and user comfort. Extracting useful knowledge from this data is challenging particularly where the underlying system is not well understood. This work demonstrates an application of Q-Analysis, a computationally simple topological approach for summarizing large sensor data sets and exploring the relationships between the different variables. The technique is applied to building energy management data from a large building that has been retrofitted with a sensor network. The technique is shown to be effective at extracting useful knowledge of the underlying system, which are highly specific and complex in nature, providing an understanding of the floor behaviors rather than the individual terminal unit performances.