The research presents an approach based on data analytics and automatic learning mechanisms integrated into one of the most widely used digital learning platforms in the world (open EDX) as a contribution to the improvement of learning processes in middle school students in municipalities of Colombia. Methodologically, according to Creswell y Plano Clark (2007), and Clements et al. (2017) a system was created to recommend educational contents adapted to the individual characteristics of students considering the limitations in the use of technologies in educational institutions; we developed interviews with teachers and focus groups with students of 10º and 11º grade; as a result, a functional architecture proposal was generated that allows the generation of initial recommendations of contents managed according to the students' performance and the characteristics of the territory.

In recent years, new trends and methodologies have emerged that greatly favor the education sector. E-learning as an alternative to regular teaching and learning processes has transformed the educational dynamics thanks to the inclusion of MOOCs, personal learning environments, allowing the educational process to be carried out at a personalized level where the focus is on learning styles and the profile of the student. This article presents a review of current works around machine learning mechanisms to make recommendations in the educational environment, where it is found that besides the discovery of the student’s learning style, it is important to know their level of knowledge and learning speed, in addition to the tools used by the student to carry out their studies. Finally, the opportunity for implementation and research of these issues in Colombia is highlighted.