Banca de QUALIFICAÇÃO: ANDRÉ GUSTAVO BOBRZYK

MINNED DATA VISUALIZATION MANAGEMENT MODEL IN VIRTUAL LEARNING ENVIRONMENTS AIMED AT THE PREDICTION OF SCHOOL RETENTION, EVASION AND DROPOUT

Learning Analytics; Dashboard; Moodle; School dropout; User experience.

The retention and school dropou of students in formal education pose important
problems for society as a whole. Dropout, in particular, determines a lower intellectual
and technical qualification of the population of a country, causing problems in social
and economic development. On the other hand, retention often causes the
impossibility of universal offer of disciplines in courses with classes that have an
excess of students and, ultimately, is the first step towards evasion. In addition, both
situations have a cost for educational institutions and, therefore, for the nation.
Currently, many courses use Virtual Learning Environments that present a significant
amount of data about student interaction. Such data can be used for the prediction of
school retention, dropout and dropout. Although there is much research in the areas of
Learning Analytics and Data Mining on the search and analysis of this data, it is
essential that VLE environments clearly present to teachers those students who are at
greater risk of dropping out or being retained. It is also accompanied that teachers can
choose the algorithms they want to use without needing deep knowledge about data
mining. Thus, the question that arises can be defined by the following sentence. "How
is it possible to build a friendly interface that allows teachers and educational
managers, from the VLE Moodle, to determine the parameters they want to use to be
alerted about students at risk and evasion, as well as to present the results in an
intelligible way and facilitated for this group of users?
In view of the above, an attempt will be made to build a plugin in the form of a Moodle
plugin, using the best data visualization and Human-computer Interaction techniques,
aiming to make the choice, filtering and presentation of data on students at risk of
retention , evasion and school dropout, intuitive and uncomplicated, even for those
teachers and educational managers who are not familiar with technological data mining
tools. The data will come from students' relaxation records in the Moodle environment
and will be pre-mined in the complementary work to this one, developed by PPG
colleague and research group on educational data mining, Pablo Oliveira, with the
guidance of Prof. Mariano Nicholas.