Application of Visual Data Mining for Visualization of UKBI Achievement Data

Abstract

The current visualization of Adaptive Indonesian Language Proficiency Test (UKBI Adaptif) results in Jambi Province is suboptimal, often relying on static, basic charts, which hinders transparency and the effective formulation of evidence-based language policies. This research aims to address this critical gap by developing an interactive, data-driven system to analyze the language proficiency profile of UKBI participants in Jambi from 2021 to 2024. The research objective is to accurately map regional competence, identify hidden patterns, and provide actionable intelligence to the Jambi Language Center. The study adopts the Visual Data Mining (VDM) methodology, integrating interactive visualization with the K-Means clustering algorithm. This method allowed for the normalization and grouping of over 10,000 participant data points, with the optimal number of clusters determined by the Silhouette Score. The research results successfully established three distinct proficiency clusters, including a "Listening Struggler Group" dominated by non-education professions, exhibiting significantly low scores in the Listening section. Furthermore, geographical analysis revealed a disparity where Jambi City—the region with the highest participation—maintained an average proficiency at the lower boundary of the Intermediate category, while smaller regions like Muaro Jambi showed higher rates of Superior and Exceptional achievement. The conclusion is that the VDM-based interactive dashboard is a validated and effective tool that successfully provides micro-level insights, supporting the strategic allocation of resources and the design of targeted intervention programs to address specific skill weaknesses, such as listening comprehension.