Deep Learning for Numeracy: A Roadmap to Asta Cita in Primary Education

The consistently low numeracy performance of primary school students, as evidenced in national and international assessments, highlights the pressing need for pedagogical reform that prioritizes meaningful and contextualized learning. This study aims to develop a conceptual framework for mathematics instruction grounded in deep learning to enhance students’ critical numeracy skills in accordance with Indonesia’s national development goals, Asta Cita. Using a systematic literature review combined with thematic analysis, the study reviewed 35 peer-reviewed international journal articles. The selected articles focused on primary students aged 7–12 and were analyzed using document analysis techniques. Braun and Clarke’s thematic analysis was applied to uncover key pedagogical approaches. Findings indicate that deep learning strategies such as real-world problem solving, project-based learning, reflective inquiry, and differentiated instruction, effectively promote higher-order thinking, problem-solving capabilities, and lifelong learning habits. These strategies also align with efforts to ensure inclusive, culturally relevant education and national autonomy. The study concludes that embedding deep learning within primary mathematics instruction not only strengthens numeracy achievement but also contributes to realizing Indonesia’s Golden Generation 2045.