Uma arquitetura para tomada de decisão pedagógica com dados integrados.

Abstract

There is a growing demand for online education and in its accomplishment play a very important role the so-called virtual learning environments, or in a more general sense the educational platforms, responsible for the provision of digital resources and support mainly to students and teachers. Teachers play highly complex roles ranging from the management of online courses to the pedagogical support offered to students, these in larger and larger classes, thus involving sophisticated decision-making tasks. To do so, these teachers increasingly require high-quality information, allowing effectiveness in their understandings in detail of what academically happens to students interacting in the online environment, and thus impacting their understandings in pedagogical decision-making processes within online learning environments. At this point, it is important to emphasize that the interactions of students in such courses generate a large amount and diversity of data, in this case, allowing the extraction of pedagogically relevant information, usually using data mining techniques or learning analytics. However, there is a significant lack in the support of teachers, since the current educational platforms present several difficulties related to the complexity in monitoring and evaluating the students involved in these courses, as well as the lack and / or difficulty in accessing relevant information that support the pedagogical decision-making process. Moreover, it is not expected that teachers master the technical tools to process all the involved data. Thus, with a view to a subsequent effective attendance to the demands of the teacher, this research focused on a problem that concerns aspects of software engineering in education, in a context of educational business intelligence, seeking to offer designers and software developers, an infrastructure necessary for the construction of solutions that meet the information demand of teachers, under the terms mentioned above. In this thesis, we propose a suitable software infrastructure for the construction of effective solutions in the provision of relevant information for supporting pedagogical decision making. In this way, it is intended to help developers in three complementary ways: (1) to favor the architectural design of the software, in order to satisfy fundamental quality requirements for this type of system, such as scalability and ease of evolution; (2) favoring the reuse of source code, together with design patterns; (3) favor integration with existing tools such as frameworks and virtual learning environments. The obtained results show that the proposed solution does, in fact, meet the desired quality requirements and facilitates integration with educational platforms, as well as favor the reuse of design patterns and development frameworks, achieving a reuse rate of about 88% . The attendances to the quality requirements were evaluated in an experiment that adopts an approach based on execution scenarios and using a real database of a course of distance information systems. The execution scenarios used are realistic and have been suggested in brainstorming meetings with experts and master’s students.