Learning analytics acts as an ecosystem of methods and techniques that continuously collect, process, report, and act on machine-readable data to enhance learning environments in higher education. Aligning with the growing need in higher education for data-informed decision-making, learning analytics has undergone significant transformation, evolving from traditional data-driven methods to advanced artificial intelligence (AI)-based approaches. While traditional approaches emphasize descriptive and diagnostic analysis, AI-based approaches focus on predictive and prescriptive insights, offering a deeper understanding of student learning outcomes that traditional methods might overlook. This study conducts a hybrid review with a primary focus on bibliometric analysis, complemented by traditional literature review methods, based on 8979 documents on learning analytics indexed in Scopus through December 2024. The results indicate a steady increase in research on learning analytics over the years, expanding across key areas such as higher education policies, social and collaborative learning analytics, student performance assessment, data mining for educational feedback, and adaptive learning. Further insights reveal a transition from traditional to AI-based approaches, reflecting a paradigm shift in how educational data is processed and utilized. Traditional learning analytics relied on descriptive and diagnostic methods to assess student performance and engagement through statistical analysis of historical data but lacked predictive capabilities and personalized interventions. In contrast, AI-driven learning analytics leverages machine learning, deep learning, and natural language processing to forecast student outcomes, identify at-risk learners, and automate assessments. This shift represents a strategic evolution in higher education, where the continuous advancement of learning analytics enhances teaching by refining instructional methods, enriches learning through personalized experiences, and strengthens decision-making with data-driven insights.