Energy efficiency and energy consumption awareness are a growing priority for many countries. Among the large variety of methods proposed by energy scientists and professionals to evaluate building energy consumption, a widely adopted approach is the energy signature. Since the energy data easily scale towards very large datasets, the problem of characterizing energy efficiency through the energy signature from these huge data collections becomes challenging. This paper presents a distributed system, named ESA, for the collection, storage, and analysis of a large amount of energy-related data to keep continuously informed users on their energy consumption and building performance. ESA exploits a Big Data approach to perform a scalable and distributed computation of the building energy signature, which is exploited to forecast the expected power consumption for given contextual conditions in a specific time period. ESA characterizes monitored buildings through direct indicators designed to (i) evaluate the efficient use of the heating system by comparing latest observations with past energy demand in the same conditions, (ii) rank the overall building performance with respect to nearby and similarly characterized buildings. Experimental results on real energy consumption data demonstrate the effectiveness and the efficiency of the proposed distributed system to provide actionable knowledge at user fingertips for actors interacting with ESA.