During the last two years remote education has taken an important place in higher education in Colombia, especially during the pandemic generated by COVID-19. Supported by Artificial Intelligence models, opinion mining processes can be automatized and applied to different text formats. For this reason, the purpose of this work is to characterize the opinions of students via the identification of aspects inside the academic environment, in order to recognize the deficiencies in their implementation in remote modality courses development, and in this way make the pertinent improvements. First, the text classification process is conducted, by identifying each of the aspects in the comments with the use of BERT, and then, an aspect-based sentiment analysis is made using PySentimiento, a sentiment analysis tool in Spanish. Finally, a comparative and descriptive analysis is made using Topical Modeling through BERTopic, in both pre-pandemic and post-pandemic periods, in order to identify the changes between these two periods.