Abstract:

Digital technology enables significant improvement of distance education (accessibility, time independence, multimedia, various methods of knowledge testing). However, the problem of assessing the quality of knowledge acquisition in online learning remains. The collected data of e-learning form the basis for the application of learning analytics to improve the quality of assessment of acquired knowledge in online learning. The use of data visualization technologies significantly increases the transparency and traceability of the results of learning analytics and thus the informative value of the success of e-learning. By designing and developing a system for visualizing e-learning data, experts are enabled to perform learning analytics procedures for individuals, groups of students, or all students for different types of knowledge reviews. By visualizing the results obtained, experts can assess the success of online education faster and better. Using technologies, experts choose the file format for exporting various e-learning data (from LMS, quizzes, etc.), but data on the results of knowledge tests are of particular importance for assessing the acquired knowledge. The e-learning data visualization system provides a graphical representation of results of knowledge testing for predefined activities (types of knowledge tests, assessment methods, passing conditions) on imported data from files for predefined knowledge tests. In order to analyze testing activities not defined in the system, the system must be customized by the system administrator or the person responsible for designing the learning analytics. The research problem relates to the possibility of automated data entry from exported files for different types of knowledge testing activities. The automation should allow that system for visualization can be adapted for different knowledge testing activities without requiring activities of the person responsible for designing learning analytics. At the beginning problem of using a system for visualization for exported files with different knowledge testing activities is defined. After that, a solution for automation in processing different types of exported data of e-learning results is presented. The solution is based on the use of metadata about knowledge testing result data in exported files with the aim of connecting the system for visualization with types of knowledge tests and knowledge test results. When importing data from the files, the structure of the dataset is analyzed, after which the expert decides in what way knowledge test activities should be connected to knowledge test results (connecting multiple tests to a group of test activities, groups of students, etc.). Metadata is generated for the imported test results so that the system for visualization recognizes the types of test activities and links them to the test results. This allows importing data from files containing data on exam results for different exam activities, without the system’s adaptation by person for designing learning analytics. The relational schema used for the described problem is presented, as well as the algorithm for importing data, creating metadata, and preparing the system for visualization to use imported data for knowledge review activities. Verification of system results with various e-learning data exports is presented, followed by an analysis of the benefits of using automation and a plan for future work.

Keywords:

System for visualization, learning analytics, exported e-learning data, adaptation.