Unveiling Numeracy Competency Domains of High School Students in Indonesia: A Clustering Analysis Approach

Authors

  • Adhi Surya Nugraha Universitas Sanata Dharma
  • Marcellinus Andy Rudhito Universitas Sanata Dharma

DOI:

https://doi.org/10.26594/jmpm.v10i2.5972

Keywords:

numeracy, student competency, national assessment, data science analysis, k-means, clustering

Abstract

This study aims to analyze high school students' numeracy achievement using a data science approach to the 2021–2023 National Assessment data. The analysis covers the domains of Algebra, Geometry, Number, and Data and Uncertainty at three ability levels: knowing, applying, and reasoning. The research methods included secondary data collection, descriptive cleaning and analysis, interdomain correlation, K-Means clustering with Davies–Bouldin Index validation, and result visualization using Python. The results show an increasing trend in national numeracy scores (49,23 to 50,76 to 55,67). The Data & Uncertainty domain reached the highest average (56,79) while having the strongest correlation with the total score. In the cognitive dimension, there was a surge in the ability to applying and reasoning in 2023. The clustering results show that the optimal configuration in 2021-2022 is 7 clusters, while in 2023 it reduces to 3 clusters. These findings suggest prioritizing strengthening data literacy and reasoning skills, along with remediation of number concepts, and targeted interventions based on provincial cluster profiles.

Downloads

Published

2025-10-17