back to the blog

Predicting KPI Improvement Factors with Machine Learning (XGBoost Classifier)

Predicting KPI Improvement Factors with Machine Learning (XGBoost Classifier)

Identifying & Predicting KPI Improvement Variables with Machine Learning (XGBoost Classifier)

Objective:

To identify factors influencing individual performance and assess the key determinants of performance improvement (KPI increase) using survey and employee data at BPJS Kesehatan (Indonesian Body of Social Security for Health) 

Challenge:

The client needed an efficient way to model individual performance increases (KPI 2022) and understand the top contributing factors affecting performance.

Approach:

  • Data Collection: Gathered survey and HR data, including metrics such as commitment, engagement, job readiness, leadership, change readiness, behaviors, values, and employee demographics.
  • Machine Learning Modeling:
    • Used the XGBoost Classifier with optimization techniques (Scale_Pos_Weight) to predict KPI improvements.
    • Conducted feature importance analysis to highlight the top 25 factors affecting individual performance.
  • Insights Development: Segmented factors by organizational, supervisor, and individual levels to identify actionable insights for each group.

Key Variables Identified:

  • Commitment: Recruitment and selection, performance management, talent management.
  • Satisfaction & Engagement: Facilities, teamwork, and communication.
  • Leadership: Visionary and effective communication.
  • Change Readiness: System thinking and adaptability.
  • Employee Characteristics: Gender, tenure, class, and past performance scores.

Tools & Techniques:

  • Data Processing: Python (pandas, NumPy)
  • Modeling: XGBoost Classifier, Feature Importance Analysis
  • Visualization: Tableau

Outcome:

  • Identified key drivers of performance improvement, enabling targeted interventions at organizational, managerial, and individual levels.
  • Created a comprehensive framework for ongoing performance monitoring and improvement planning.
  • Delivered actionable recommendations that enhanced employee productivity and organizational KPI alignment.

Stacks

  • Machine Learning & Predictive Analytics
  • Google Cloud Platform (GCP)
  • Data Visualization & Business Intelligence 
  • Pricing Models & Market Analysis
  • Python, SQL, Oracle, Django
  • Streamlit, Tableau