By combining Artificial Intelligence and Data Mining, complex domain problems can be analyzed. The knowledge of energy consumers’ profile has been an important tool when making decisions in the energy sectors. The aim of this paper is to explain and evaluate the use of different Data Mining techniques and learn which strategies are more adequate to analyze consumer profiles. By using smart meters, raw data can be easily collected and stored in a database. Extensive volumes of data is generated and optimal efficiency in data management is crucial. Such data have a time-series notion typically consisting of consumers’ energy usage measurements over a time interval together with a detailed level of consumer profile data. Data preprocessing is vital for reliable insights and an appropriate choice of data transformation provides better outcomes in model construction. In this paper, anonymized electricity power consumption data will be analyzed. Data preprocessing using weekly time intervals is proposed together with an analytical model that consolidates observations of weekly time series by consumer. Following the generation of consumers’ weekly energy usage, a set of typical weeks will be found and compositional data analysis will be applied. The performance of the proposed model will be evaluated and the results will be discussed.