Itaipu, the greatest hydroelectric power plant in operation in the world, has more than 2,200 monitoring instruments, which have been storing readings in databases for more than 30 years. The large data sets of high dimensionality and the large amount of records inserted into databases are non-trivial problems when conducting a search for "knowledge" through the data. This paper introduces a study using Visual Data Mining (VDM) algorithms integrating Data Mining (DM) techniques with Visualization of Information (VI) techniques to analyze the data collected at Itaipu dam. The main objective was to establish relationships between the variables in order to detect undesirable failures that can compromise the security and integrity of the dam. More information may be more easily extracted when different techniques of Visualization of Information, together with techniques of Data Mining, are applied for data analysis. The visual analysis of the data has proved efficient in detecting patterns of anomalies, and thus it can be considered a valuable tool to support decision making.
KDD process; Information visualization; Data mining; Visual data mining; Dam monitoring
