Detecting students-at-risk in computer programming classes with learning analytics from students’ digital footprints

Abstract

Different sources of data about students, ranging from static demographics to dynamic behavior logs, can be harnessed from a variety sources at Higher Education Institutions. Combining these assembles a rich digital footprint for students, which can enable institutions to better understand student behaviour and to better prepare for guiding students towards reaching their academic potential. This paper presents a new research methodology to automatically detect students "at-risk" of failing an assignment in computer programming modules (courses) and to simultaneously support adaptive feedback. By leveraging historical student data, we built predictive models using students' offline (static) information including student characteristics and demographics, and online (dynamic) resources using programming and behaviour activity logs. Predictions are generated weekly during semester. Overall, the predictive and personalised feedback helped to reduce the gap between the lower and higher-performing students. Furthermore, students praised the prediction and the personalised feedback, conveying strong recommendations for future students to use the system. We also found that students who followed their personalised guidance and recommendations performed better in examinations.

Predicting Students Success in Blended Learning—Evaluating Different Interactions Inside Learning Management Systems

“…The results have shown that the techniques used in the study can identify students with the risk of failing, where the best results were achieved using the Support Vector Machine (SVM) algorithm. Azcona et al [26] present a research methodology to detect at-risk students in computer programming courses too. The authors provide adaptive feedback to students based on weekly generated predictions.…”

Preface to the special issue on learning analytics and personalised support across spaces

“…Detecting students-at-risk in computer programming classes with learning analytics from students' digital footprints (Azcona et al 2019) Detecting at-risk students Adapted group formation Behaviour in multiple sub-spaces in a MOOC and a LMS mostly conducted manually, this paper also points at the potential of using emerging multimodal sensors to study multiple dimensions related to knowledge building (e.g. emotional and cognitive) in both physical and digital spaces.…”

“…In the paper titled Detecting Students-at-Risk in Computer Programming Classes with Learning Analytics from Students' Digital Footprints, Azcona et al (2019) propose a novel methodology to automatically detect students at risk of failing an assignment in computer programming courses and to provide weekly adaptive feedback for both students and lecturers in large classes to react upon. Authors embraced the spirit of learning across spaces by creating models that involve sets of static information about the students, such as demographics and prior academic history, and behavioural logs captured from two different digital spaces: a specialised programming tool and the institutional learning management system (LMS).…”

Putting It All Together: Combining Learning Analytics Methods and Data Sources to Understand Students’ Approaches to Learning Programming

“…The analysis of learners' data from online learning environments on its own may fail to account for the full breadth of students' learning actions if part of their learning process takes place outside this scope, such as in Integrated Development Environments (IDEs). With this in mind, authors have analyzed the data collected from auxiliary systems that support programming students in their learning process (e.g., automated assessment systems for programming assignments), to examine how different factors affect students' outcomes and programming mastery, such as code quality, code size, and usage frequency [18][19][20]. In this study, we argue that a holistic approach that is based on the analysis of data originating from several relevant sources to programming education would provide a more accurate view of students' learning, considering how the temporality of the learning tactics, the interconnectedness of events, and the combination of various learning strategies may influence the outcome of the learning process [21].…”