Artificial Intelligence is one of the sciences that nowadays has increased its participation in multiple contexts [Malik et al., 2018] including medicine, agronomy, productive systems, education, among others. In this document it's presented the creation process of an artificial intelligence based - emotion detection system that is going to make part of a platform which main purpose is to give support in a proactive way to the students of Pontificia Universidad Javeriana to reduce the student’s desertion rate. The model is created using the CRISP-DM for data mining projects methodology, obtaining as result an API/REST definition that makes emotions predictions using as input data an image of the student's face.