The amount of data collected by the SCADA (supervisory control and data acquisition) of an urban water supply system is sometimes difficult to process. A multivariate statistical technique, Principal Component Analysis (PCA) is presented in this paper, which processes this data, simplifying and synthesizing the most significant information. This technique extracts new variables, principal components (PC), that explain the behaviour of injected flow. Multivariate control charts to detect outliers show higher sensitivity than those generated with traditional univariate statistical methods.