Plastic color mismatch causes identification using OLAP and data mining

Most companies have their data stored electronically. The appropriate processing of this data can quickly identify issues leading to process improvement and cost reduction. However, the manipulation of the data stored in the companies' repositories is not trivial for decision support. In this paper we propose the use of Excel spreadsheets for the detection of parameters that may cause the plastic color rejects at one of SABIC Innovative Plastic's plants. The entire complex data processing can be executed directly by the manufacturing engineer in a user-friendly environment by using the combination of OLAP and data mining. The results show that it is possible to reduce the ratio of the parameters that may be the cause of color rejection problems to less than a percent. This provides the opportunity of studying the scientific reasons for it, thereby leading to improvements in first time color pass production.