Development of an analytical dashboard for predicting attrition and talent management using data mining and machine learning on personnel data
Development of an analytical dashboard for predicting attrition and talent management using data mining and machine learning on personnel data
Najmeh Nazer Mozaffari1
1) Administrative Admissions Expert, Human Resources, Shiraz Municipality
IOI Code
:
AAHF-ABBCEC
Publication :
National Conference on Architecture and Art Technologies, Smart City, Soft and Creative Industries and Technological Culture(icact.ir)
Abstract :
Key talent attrition, as a major challenge for organizations, brings heavy financial and operational costs. Traditional approaches to attrition management are often reactive and imprecise. The new paradigm of “human resource analytics” using data mining and machine learning enables scientific prediction and timely intervention. The aim of this research was to design and develop an operational framework including an analytical dashboard for predicting attrition using machine learning on historical personnel data, in order to transform the data into actionable insights for managers. This applied-developmental research was conducted following the CRISP-DM framework and on five-year historical data of 3800 employees of an Iranian organization. After integrating data from different sources and preprocessing, the performance of five machine learning algorithms was compared and the best model was interpreted using the SHAP technique. The outputs were implemented in an interactive web dashboard. The XGBoost algorithm performed best with remarkable accuracy (AUC=0.927). The most important predictors of attrition were distance from last promotion, average performance score, and negative affect score. The analyses identified four attrition patterns: “stagnant ambitious,” “newly dissatisfied,” “financially stressed,” and “organizational isolated.” The simulations showed that targeted interventions (such as promotions) could reduce attrition by 42%. The research demonstrates the high efficiency of machine learning-based approaches to predict attrition and provide actionable insights. The developed dashboard, as a decision support system, enables prediction, root cause analysis, and simulation of interventions, paving the way towards evidence-based and proactive talent management.
Keywords : attrition prediction talent management machine learning analytical dashboard human resource analytics