the importance of sensor validation
Water urban environments may have hundreds or thousands of sensors which produces vast amounts of data. Nevertheless, if the data quality of sensor is poor due to errors (for example, outliers, values out of range, flat values…), the data is rendered useless because it may lead to wrong decision-making. Therefore, an essential part of working with sensors is to apply methods to ensure quality of raw data. AI techniques plays a vital role in this field contributing to more robust and accurate data. SCOREwater applies univariate and multivariate analysis based on Machine Learning and Data Science to pre-process RAW data ensuring data quality for future decision-making.
The picture below represents part of the analysis done during the Machine Learning process. It is used to visually validate that one hypothesis can be viable (checking visually that new features separate the classes)
Read more about the SCOREwater project in the year 1 overview (PDF): https://bit.ly/SCOREwater1year