Unlocking Workforce Insights: Exploratory Data Analysis for Strategic Workforce Planning in State Universities and Colleges (SUCs)

Abstract

Workforce decisions should be supported by workforce data analytics rather than relying solely on intuition.  This can now be addressed with the use of IT-based techniques such as workforce analytics, which helps to transform raw data into insightful and quantifiable results. Relative to this, this paper looks into exploratory data analysis (EDA) to gain significant insights. The study employs a quantitative approach using sample data from an SUC in the Bicol Region, Philippines. EDA is conducted within the Jupyter Notebook employing descriptive statistics, Latent Dirichlet Allocation (LDA), Jaccard Similarity, and Auto-Regressive Integrated Moving Average (ARIMA) for predictive modeling. The results of the analysis reveal insights across various dimensions such as distributions across age, gender, civil status, employment status, educational levels, length of service, and even positions that will be vacated in the succeeding years.  Notably, the LDA approach uncovers the skills and expertise of the current workforce, highlighting dominant expertise such as Information Technology and Nursing within the administrative and academic cluster, respectively. Jaccard similarity result suggests that there is a certain level of alignment and gap between the expertise possessed by the current workforce with the expertise required by the Commission on Higher Education (CHED).  Forecasts derived through ARIMA modeling project a student enrollment for the next semester, with corresponding prediction accuracy metrics such as Mean Absolute Error, Mean Squared Error, and Root Mean Squared Error. Finally, the analysis concludes with projecting for faculty workforce demand to accommodate future student enrolment, thus offering a strategic approach to workforce planning.

Keywords: Auto-Regressive Integrated Moving Average (ARIMA), exploratory data analysis, Jaccard Similarity, strategic workforce planning, SUCs, Latent Dirichlet Allocation (LDA)