Applied Statistics in Industry: Defining an Appropriate Target Variable and Analysing Factors Affecting Aluminium Ingot Quality

Methods from the field of data science are becoming increasingly important in various areas of the economy. One of the main applications in industry is the continuous improvement of product quality, which is becoming increasingly challenging due to growing quality demands. A variety of statistical and machine learning methods are available for this purpose. However, to use these methods for a specific application, it is necessary to define an appropriate target variable. This can be a significant challenge due to the extensive product portfolio. The target variable must cover many cases and remain comparable for all cases under consideration. Limiting it to a single case would often result in a database too small for any analysis. This work aims at improving the quality of aluminium plates for the aerospace industry by analysing casting process data. The focus is on the development of a target value for this issue. Additionally, a brief outlook is given on the subsequent analysis of possible influencing variables and the development of a predictive model of product quality.