Discovering unexpected information using a building energy visualization tool

In this paper we present a 3D visualization tool developed to gain insight about buildings energy consumption. We will focus on the usage of this software to extract information from a raw dataset, and more especially how unexpected features can be discovered. In this paper, we will present some related work on data visualization, our visualization tool and finally, two case studies of data analysis using visualization. This work is developed for the RIDER project (Research for IT Drive efficiencies), composed by several French companies. The goal of this project is to develop a new platform to manage buildings energy. Smart buildings are already managed by BMS (Building Management System) devices. These IT (Information Technology) equipments are used to analyze data from sensors, and then apply some action on the building. These systems are only reactive devices and they are used only to improve immediately buildings environment. Missing features of the actual BMS have been identified and components developed by this project try to solve these gaps. RIDER green box is a more complex BMS, it is composed by a set of predictive components developed to analyze and improve building usage. An example of improvement can be to increase temperature of a meeting room before the arrival of room's users. RIDER is composed by a data mining tool to dig into the large dataset provided by sensors, a model component developed to create building model with a reduced set of parameters, a decisional component develop to improve rules of building management based on user behavior, a visualization component and finally some external components like weather station or calendars. The visualization engine allows to visualize dataset in an understandable way. We use a particle paradigm for