Analyzing historical data of industrial Cleaning-in-place (CIP) operations is essential to avoid potential operation failures, but is usually not done. This paper presents a three-level framework of analysis based on the CIP case of a brewery fermenter to describe how to analyze the historical data in steps for detecting anomalies. In the first level, the system is assessed before cleaning to ensure that the selected recipe and system are able to accomplish the task. In the second level, a multiway principal component analysis (MPCA) algorithm is applied to monitor the process variables on-line or post cleaning, with the purpose of locally detecting the anomalies and explaining the potential causes of the anomalous event. The third level analysis is performed after cleaning to evaluate the cleaning results. The implementation of the analysis framework has significant potential to automatically detect deviations and anomalies in future CIP cycles and to optimize the cleaning process.