Clustering Residential Energy Consumption in Mediterranean Climate Cities

Residential energy consumption (REC) is a common parameter developing sustainability indicator systems. Deeper knowledge on the nature of its spatial distribution is a key aspect for future urban and infrastructure planning as well as for the implementation of sustainable cities models based on distributed renewable energies production and storage. This scenario challenges us to develop an information model that allows exposing the predictive parameters and its resulting effect on the likely energy consumption. The main objective is to develop a methodological contribution to provide that information model. The relevant parameters for the REC are assessed and applied to a case study focused on the whole set of municipalities across Andalusia. To conduct the required geostatistical and energy analyses for mapping the potential REC at an urban and regional scale, open datasets are used to obtain the parameters involved in energy demand and consumption. Multiple nature parameters considered are grouped into three basic categories: location, occupation, and buildings. Therefore, with the selected data and parameters, a geostatistical model is developed in a Geographic Information System (GIS). Different stages of acquisition, preparation, exploration, and data modelling are described for this study along a Data-Mining or Knowledge Discovery in Databases (KDD) data science process that allows observing patterns and trends. The results of the study provide a geographic segmentation or cluster definition among similar or peer municipalities using a density-based clustering algorithm on the set of predicting REC parameters stablished for the array of towns of Andalusia.