There are several ways to implement dashboards in Python. Based on use cases from the area "Students Advice" an overview should be created, which allows a comparison and evaluation of the tools. The topic includes a general research on the possibilities with Python and the specification of use cases and their validation as well as the arrangement of evaluation criteria.
Students achieve different performances during their studies. Though most of them enroll in all courses of the first semester, some of them choose not to seat for the exam; those who take the exam get different marks. After the first semester, behaviors differ more due to courses that students repeat.
The aim of this investigation is to find out, whether typical behaviors with respect to performance emerge with the use of clustering algorithms. A series of clustering will be conducted to account for the students who drop out over the semesters. Clustering 1 would take into account only data from the first semester, clustering 2 would take into account data from the first and second semester, and so on. A visualization that shows how students move from one cluster to the next one in two consecutive semesters should be proposed. Do clusters separate students who drop out and students who graduate?
In order to be able to offer targeted help in the context of study advisory services, it is necessary to know as accurately as possible whether a student is in difficulties and whether there is a risk that he/she may drop out of the course. In order to predict possible drop-outs, different Auto-ML toolkits will be used and their performance as well as advantages and disadvantages evaluated and compared. Which tools achieve the best accuracy with which algorithms?
Not all students follow the module handbook and take the modules in the recommended semester. For example, if you have little time due to family or main job, you will need longer to complete your studies. Can certain successful paths be identified? At what point are you on the wrong path? Which visualizations support the analysis of sequential patterns?
In this investigation, such questions should be tackled beginning with a single degree program.
Not all students pass all courses of the first semester in their first semester of study. A preliminary clustering of graduates in a degree program shows that student who follow the curriculum in each semester usually get better marks and finish their studies in the designated time. What for students who fail courses? Should they repeat them as soon as possible? Is there any optimal time-span to wait? Is this optimal time-span different for each course?
In this investigation, such questions should be tackled beginning with a single degree program and courses of the first semester.