Archives

  • Journal of Data Science Applications.
    Vol. 1 No. 1 (2025)

    This issue of Jurnal Aplikasi Sains Data (JASID) features research on diverse applications of data science methodologies. The first article compares ARIMA and SARIMA models for forecasting non-oil and gas exports in East Java. Next, a content-based filtering approach is applied to a tourist destination recommendation system in Central Java. Another study evaluates the performance of stochastic gradient descent and adaptive moment estimation optimizers in audio-based emotion classification using convolutional neural networks. Additionally, a fuzzy inference system has been developed for the quality assessment of formula milk in a stunting prevention program. Lastly, the clustering of diabetes risk factors is performed using the K-Prototypes Clustermix algorithm. Collectively, these studies demonstrate the relevance of data science techniques in economic forecasting, healthcare, machine learning, and recommendation systems.

  • Journal of Data Science Applications.
    Vol. 1 No. 2 (2025)

    Journal Aplikasi Sains Data (JASID)

    Vol. 1, No. 2, 2025

    This issue showcases five applied data-science studies that translate methods into impact across health, safety, education, and regional development. The contributions span modern machine learning, classical statistics, and interactive analytics each grounded in real Indonesian contexts and 2024–2025 data.

     

     

     

    Highlights of the Issue

    • CNN for Web-Based Indonesian–to–Sign-Language Translation
      A prototype pipeline leveraging convolutional neural networks delivers real-time sign rendering from Indonesian text, pointing toward inclusive, accessible web interfaces for the deaf community.

    • Wilcoxon Analysis of Infant Malnutrition Pre/Post COVID-19 in North Sumatra
      A rigorous nonparametric assessment quantifies shifts in malnutrition cases across pandemic phases, offering statistically defensible evidence for post-pandemic policy targeting.

    • Feature-Importance–Guided Ensemble for Thyroid Cancer Recurrence
      An interpretable ensemble framework prioritizes clinically salient predictors to improve recurrence risk stratification in differentiated thyroid cancer.

    • XGBoost for 2024 Fire-Risk Classification in Surabaya with Streamlit Maps
      Gradient-boosted models classify urban fire risk and are deployed via an interactive Streamlit dashboard, enabling spatial exploration for first responders and city planners.

    • K-Means Clustering of East Java Regencies/Cities by 2024 HDI Indicators
      Unsupervised segmentation reveals development archetypes, supporting differentiated regional interventions and benchmarking.

    Editorial Note
    Collectively, these papers demonstrate how methodological rigor, CNNs, XGBoost, ensembles, nonparametric inference, and clustering can be paired with usability (web apps, dashboards) to inform decisions. We hope this issue serves researchers, practitioners, and policymakers seeking reproducible, actionable data-science solutions for Indonesia.