From Struggle to Success: A Learning Analytics Pilot at Mamun University
Sanat Chuponov, Mamun University
This study examines a learning-analytics pilot at Mamun University, conducted in partnership with Queen Mary University of London and New Uzbekistan University. The project began with an undergraduate pilot using two semester groups, but technical problems with Moodle and weak coordination between lecture and practical teaching meant that the pilot was paused mid-semester.
The project was subsequently redesigned and implemented with a smaller postgraduate group. With Moodle ready from the beginning of the semester and a single instructor responsible for the course, engagement data remained closely connected to teaching throughout the pilot. Five engagement channels were monitored: attendance, formative assessments, weekly experiential activities, problem sets and office-hours attendance.
The postgraduate results were encouraging. Students recorded high lecture and practical attendance, strong completion of problem sets and an overall engagement score with a median of 3.54. Most notably, engagement was positively associated with final module marks, with a correlation of 0.668 and an R² of 0.447. Formative assessment showed the strongest reported relationship with final marks among the engagement components, with a correlation of 0.592.
Student feedback was also strongly positive. Among 19 survey respondents, 94.8% were satisfied overall with how engagement monitoring was used, while 89% felt more accountable for their learning and 89% felt motivated to participate more actively. Students generally viewed the process as professional, respectful of privacy and supportive of changes in how they studied and managed their time.
Key findings
Strong engagement in the postgraduate pilot: Lecture attendance averaged 86%, practical attendance 85%, formative-assessment engagement 65%, and problem-set completion 80%. Weekly-question completion averaged 36%, while office-hours attendance was 4%.
Positive engagement–performance relationship: Engagement was positively correlated with final module marks (r = 0.668), with engagement alone explaining 44.7% of the variance in final marks.
Formative assessment was particularly important: Formative assessment had a correlation of r = 0.592 with final marks and showed the strongest relationship among the engagement components reported.
Student perceptions were highly positive: 94.8% of survey respondents were satisfied with the use of engagement monitoring, 84% were comfortable with and trusted the process, and 84% felt respected.
Implementation conditions mattered: The comparison between the two attempts suggests that platform readiness and coordination between teaching staff were important conditions for the pilot. The undergraduate attempt was paused after technical delays and weak coordination, whereas the postgraduate implementation had the platform ready from week one and teaching and engagement data remained aligned.
Limitations and implications
The pilot involved a small postgraduate cohort of 26 students, with 19 responding to the end-of-module survey. The undergraduate attempt did not produce the expected results and was paused because of technical and coordination problems. The report therefore presents the postgraduate findings as encouraging rather than establishing that learning analytics will produce the same results across larger or different cohorts.
Overall conclusion
The pilot suggests that learning analytics can provide a useful structure for supporting student engagement when the underlying technology and teaching arrangements are in place. The postgraduate cohort showed a clear positive relationship between engagement and final marks, alongside high levels of student satisfaction. At the same time, the comparison between the two attempts demonstrates that successful learning analytics depends not only on the framework itself, but also on reliable technology and effective coordination between teaching and engagement monitoring.