In an educational context characterized by rapid digital transformation, learning analytics offers strong potential to personalize instruction and enhance teaching practices. However, in Moroccan pioneer colleges, this potential remains largely underexploited. This study aims to design a computational model that transforms students’ digital traces into interpretable and actionable insights to support data-driven pedagogy accessible to non-technical teachers. Grounded in Learning Analytics, pedagogical visualization, and educational artificial intelligence, the proposed model is organized into five interconnected modules. These modules process learner interaction data from educational platforms to generate key indicators such as engagement, persistence, and academic achievement. The outputs are presented through an intuitive and interactive dashboard designed to help teachers visualize and interpret learning behaviors effectively. The proposed model is designed to translate raw data into meaningful pedagogical indicators and visualizations, enabling both individual and group-level analysis. It is intended to support timely instructional decision-making through functionalities such as student grouping, disengagement alerts, and automated recommendations for differentiated instruction, all adapted to the Moroccan secondary education context and designed for teachers with limited data literacy. This study contributes a context-sensitive framework that facilitates the integration of learning analytics in classroom practice. By providing simple indicators and visual feedback, it promotes evidence-based decision-making and pedagogical differentiation. As a theoretical and exploratory model, it still requires empirical validation and teacher training for effective implementation.