SAFARI-P: Swahili-Focused Adaptive Framework for Accelerated Reinforcement in Intelligent Python Education
Vasu Jindal · 2025
This paper introduces SAFARI-P (Swahili-Focused Adaptive Framework for Accelerated Reinforcement in Intelligent Python Education), an innovative system that integrates Generative Adversarial Networks (GANs) and a Multi-Dimensional Learning Confidence (MDLC) system for Python programming education in Swahili. The framework comprises three key components: (1) an Adversarial Code Generation System (ACGS) for creating culturally relevant code snippets, (2) an MDLC Assessment Module that uses a tripartite confidence matrix system to evaluate concept understanding, problem-solving patterns, and code quality, and (3) a Cultural Context Integration Engine for incorporating local elements. By simultaneously tracking technical proficiency and cultural understanding through confidence matrices, SAFARI-P provides personalized learning paths that ensure students build strong foundational knowledge while maintaining cultural relevance. In a 16-week study conducted across Kenya, Tanzania, and Uganda, this mathematically rigorous approach to confidence tracking led to significant improvements: a 27% increase in Python proficiency (p < 0.001) and a 32.6% improvement in problem-solving efficiency (p < 0.001).