The water distribution system constantly aims at improving and efficiently distributing water to the city. Thus, understanding the nature of irregularities that may interrupt or exacerbate the service is at the core of their business model. The detection of technical and non-technical losses allows water companies to improve the sustainability and affordability of the service. Anomaly detection in water consumption is at present a challenging task. Manual inspection of data is tedious and requires a large workforce. Fortunately, the sector may benefit from automatized and intelligent workflows to reduce the amount of time required to identify abnormal water consumption. The aim of this research work is to develop a methodology to detect anomalies and irregular patterns of water consumption. We propose the use of algorithms of different nature that approach the problem of anomaly detection from different perspectives that go from searching deviations from typical behavior to identification of anomalous pattern changes in prolonged periods of time. The experiments reveal that different approaches to the problem of anomaly detection provide complementary clues to contextualize household water consumption. In addition, all the information extracted from each approach can be used in conjunction to provide insights for decision-making