Process data acquired during manufacturing processes is time-series data, such as characteristic values and process parameters obtained from sensors and measuring instruments. Understanding temporal trends is important for identifying factors affecting characteristics. In this presentation, we will introduce examples of exploratory analysis using JMP for input information and characteristic values (output) obtained every second in a continuous flow manufacturing process, such as environmental conditions, materials in the machine, temperature, and pressure. Specifically, we visualized time trends using Graph Builder, reduced the dimensionality of multivariate data using principal component analysis, and identified process change points using partitioning (regression tree analysis). These analyses demonstrated the potential for deeper understanding of process states and their usefulness in identifying factors based on temporal trends. Furthermore, we will discuss the importance of incorporating temporal information when building predictive models such as regression.
