In recent years, Governments of many countries had promoted higher education in exclusive areas and disciplines to improve the quality of inhabitant's life, economy, and public management. However, there is no an effective way to help governments to verify if scientific works and research centers are aligned to a country priority research area. A dataset of 4552 scientific works associated to 29 research areas in Ecuador was collected from a bibliographic database SCOPUS. This work is focused on detection of the most important Ecuadorian research areas based on scientific collaboration in scientific publications in such areas through a proposed methodology. The methodology is subject to Social Network Analysis and Natural Language Processing to identify collaboration communities and the most relevant topics on densely connected communities. Finally, an exploratory analysis of a case study thought this methodology is presented to demonstrate that is possible to know which scientific areas are the most collaborative and which topics were the most popular in Ecuador.