The study aims to understand how collected data from the Assembly Wire Bond manufacturing line and enterprise systems can lead to insights regarding the relationship of overtime rate to manufacturing KPIs: output attainment, utilization, and absenteeism. Using CRISP-DM as the data mining methodology, approximately 3.8 million records for attendance, 25.7 million for lot transactions, and 546.2 million for machine utilization were extracted, transformed, and summarized into weekly buckets, resulting in three hundred fifteen (315) samples from the five (5) selected packages, representing 97% of the total volume. The samples were subjected to Multiple Linear Regression and Pearson Correlation to determine whether there was a significant relationship between Overtime Rate and the manufacturing KPIs. Results show that Overtime Rate exhibits a statistically significant relationship with all the variables (P < 0.05) for the overall setup. Violations of the regression assumptions are probably due to model misspecification: other variables can more effectively explain the behavior of the overtime rate, as indicated by the low R-squared values. In contrast, only selected variables are considered significant across the results by package. These differences can be attributed to the different environmental conditions of semiconductor processes and manufacturing lines.