Energy Analytics in Public Buildings using Interactive Histograms

Abstract

In this paper we propose a visual analytics approach based on data cube methods to provide an insightful analysis of how energy is being used in a group of public buildings according to many different factors. The analysis is done by means of a web-based visual interface featuring “live” coordinated views – histograms – that show the distribution of demand data, according to different attributes, under different scenarios defined by user-driven filters on these attributes. We use the crossfilter.js library to achieve real-time computation of data cube aggregations for constantly changing user-defined filters, resulting in a fluid visualization of demand parameters (active power, power factor, total harmonic distorsion, etc.) aggregated according to many different factors or dimensions such as time (hour, day of week, month, etc.), building or environment (outside temperature)