Analisis Cost-Effectiveness Gaji Karyawan Menggunakan Exploratory Data Analysis (EDA)

Abstract

This research focuses on auditing the Cost-Effectiveness of the payroll system at PT XYZ to detect inefficiencies in human resource budget allocation. Unlike conventional performance evaluations, this study applies an Exploratory Data Analysis (EDA) approach to diagnose operational data anomalies, such as extreme hourly wage cost distortions caused by incomplete timesheet data. Data preprocessing was conducted using Python by applying median imputation techniques to handle outliers in working hours without reducing data volume. Advanced analysis was performed using SQL to formulate a Payroll Effectiveness Score (PES) with a dynamic weighting scheme for salary, working hour productivity, and tenure variables. Interactive visualization via Tableau revealed significant efficiency disparities among business units, where major operational branches recorded the lowest effectiveness scores. Based on the 0.2 quantile threshold, this study successfully isolated specifically 17.61% of the total employee population as priority candidates for rationalization (layoff). These empirical findings are recommended as a strategic foundation for management in restructuring salary budgets based on actual productivity.