PENGEMBANGAN GROW-SD BERBASIS LEARNING ANALYTICS, ADAPTIVE INTELLIGENCE, DAN DEVELOPMENTAL ASSESSMENT FRAMEWORK DI SEKOLAH DASAR

Fragmented and periodic documentation of elementary school students' learning progress frequently hinders educators from establishing longitudinal developmental trajectories and executing early pedagogical interventions. This persistent gap necessitates an integrated instructional system capable of transforming formative records into actionable pedagogical insights. This study aims to develop and validate GROW-SD (Growth Monitoring System), a web-based educational platform that synthesizes learning analytics and adaptive intelligence within a developmental assessment framework to guide data-driven teaching decisions in elementary schools. The research methodology integrates Design-Based Research principles with the ADDIE instructional engineering model across the analysis, design, and development phases. Needs analysis data were collected from 15 elementary schools in East Java through teacher surveys, classroom observations, and focus group discussions with teachers and school principals, followed by comprehensive functional testing and expert validation involving six specialists. The diagnostic assessment revealed that 93.3% of teachers required automated developmental dashboards, 90.0% demanded early risk identification mechanisms, and 86.7% needed adaptive intervention suggestions. GROW-SD emerged as a functional minimum viable product featuring a five-layer system architecture, 16 multi-dimensional developmental indicators, and 28 rule-based recommendation guidelines. Internal functional evaluations confirmed that all 38 operational test scenarios executed successfully with zero critical errors, while expert reviews achieved an overall feasibility score of 92.4% (highly feasible). The study concludes that GROW-SD possesses robust conceptual, technological, and interface readiness for restricted field trials, although its sustainable empirical efficacy in accelerating academic progress warrants extended longitudinal validation. ABSTRAK Catatan pembelajaran siswa sekolah dasar yang terfragmentasi dan periodik menyulitkan pendidik dalam memantau lintasan perkembangan belajar secara longitudinal serta menghambat pengambilan intervensi sejak dini. Kondisi tersebut menuntut hadirnya sistem informasi terpadu yang mampu mengolah data asesmen formatif menjadi wawasan pedagogis yang operasional bagi guru. Penelitian ini bertujuan untuk mengembangkan dan memvalidasi GROW-SD (Growth Monitoring System), yaitu sistem berbasis web yang mengintegrasikan learning analytics, adaptive intelligence, dan developmental assessment framework untuk mendukung pengambilan keputusan pembelajaran di sekolah dasar. Metode penelitian yang diterapkan memadukan pendekatan Research and Development (R&D) berbasis Design-Based Research dengan model ADDIE pada tahapan analisis, perancangan, dan pengembangan prototipe. Pengumpulan data kebutuhan melibatkan 15 sekolah dasar di Jawa Timur melalui survei terhadap guru, observasi kelas, dan diskusi kelompok terfokus bersama guru serta kepala sekolah, yang dilanjutkan dengan pengujian fungsional dan validasi oleh enam orang ahli. Hasil analisis kebutuhan menunjukkan tingginya urgensi dasbor perkembangan terpadu (93,3%), deteksi dini risiko belajar (90,0%), dan rekomendasi intervensi adaptif (86,7%). Prototipe GROW-SD berhasil dikembangkan sebagai minimum viable product yang memuat arsitektur lima lapis, 16 indikator perkembangan, dan 28 aturan inferensi adaptif berbasis aturan. Pengujian fungsional internal membuktikan seluruh 38 skenario berhasil dijalankan tanpa galat kritis, sementara uji validasi ahli memperoleh skor kelayakan sebesar 92,4% yang berkategori sangat layak. Penelitian ini menyimpulkan bahwa GROW-SD memenuhi kelayakan konseptual, teknis, dan antarmuka untuk diuji coba secara terbatas, meskipun efektivitas jangka panjangnya terhadap peningkatan hasil belajar masih memerlukan pembuktian melalui implementasi lapangan longitudinal.