In this chapter, we present a project called RIDER, for Research for IT Driver EneRgy efficiencies. We are developing a green box composed by a set of IT components; able to generate efficient predictive model and building optimization. The main part of the green box is a pivot model managed by a rule engine and a standard BMS with a specific input/output protocol. Modules can extend the standard green box offering system/user advices, proposing improvements on data model and predicting building behavior. This chapter introduces the RIDER green box and details the visualization module, which is used to improve model by using user knowledge.