Student performance analytics
See Student Performance While You Can Still Change It
End-of-semester results are a record, not a tool. Performance analytics is the same information arriving in week four, at the level of the concept that broke, for the people who can still do something about it.
Course analytics — Semester 4, Database Systems
- Students
- 186
- Weak concepts
- 4
- Avg. mastery
- 67%
-
Relational modelling Strong
-
Query optimisation On track
-
Transactions and locking Needs attention
Illustrative dashboard. Not customer data.
What this actually means
Reporting describes the past. Analytics changes the present.
Most institutions already produce reports: marks by course, pass percentages, attendance registers. Those are accurate and they arrive too late to act on. Performance analytics is a working dashboard, refreshed as assessments and attendance land, built so that every number on it points at a next step.
A mark is an outcome
Sixty-one percent tells you the result. It does not say which four questions were missed, or that three of them tested the same concept.
A concept is a cause
Because every question is tagged with what it tests, a result decomposes into concepts. That is the level at which teaching can respond.
A trend is a decision
The same concept measured twice shows whether last month's intervention worked. Without the second measurement, nobody knows.
The views
The same data, cut for the person looking at it
A student, a lecturer, a head of department and a placement team need four different answers from one set of records. Each role sees the cut that matches what they are responsible for, and nothing beyond it.
-
Student
My progress
Mastery by concept across current courses, what improved since the last assessment, and the one or two topics to work on next.
-
Faculty
This course
Distribution across the class, the concepts the cohort failed together, and the students whose trend is falling rather than flat.
-
Head of department
Across courses
Performance by course and by semester, with participation and coverage beside it so a weak result can be read in context.
-
Management
Across the institution
Programme-level trend over time, comparable between departments, without waiting for the examination cycle to close.
-
Placement
Readiness
Performance mapped to the competencies employers screen for, so a placement conversation starts from evidence rather than a transcript.
-
Advisors
My students
A caseload view of assigned students with their current standing, recent movement and the flags raised since the last meeting.
The metrics
What each number on the dashboard is made of
A metric nobody can define is a metric nobody trusts. These are the four that carry most of the weight, and what each one is computed from.
Concept mastery
The share of questions tagged to a concept that a student answers correctly, across every assessment that touched it. It moves when new evidence arrives, so it is a current position rather than a historical average.
Participation
Assessments attempted against assessments set, and attendance against sessions held. Reported separately from performance, because a student who is absent and a student who is present and struggling need different responses.
Movement
The change in a student's mastery on the same concept between two measurements. This is the number that tells you whether teaching or remediation changed anything, and it is the one most reporting leaves out.
Cohort spread
How wide the distribution is on a concept, not just where the average sits. A course with a healthy mean and a long tail is a different teaching problem from one where everybody is equally adrift.
Analytics describes what happened and how it is moving. It does not decide who passes, and it does not produce a score for a student that nobody can explain. How these signals turn into an early flag.
Questions
What departments ask about the numbers
Where does the data come from?
Assessments taken on the platform, attendance, and submissions. Where a college runs an academic ERP, student records, programmes and enrolment sync across so the cohorts match your own structure. We would rather integrate with the system of record than ask a department to key data twice.
How is this different from the reports our ERP already produces?
An ERP report is usually accurate, periodic and at the level of a mark. This is concept-level, continuous, and built around movement between two measurements. The two are complementary: the ERP remains the record, and the analytics layer is what you look at during the term.
Can students see their own analytics?
Yes, scoped to themselves. A student sees their own mastery by concept, what has improved and what to work on next. They do not see other students, and they do not see the cohort comparisons a head of department works with.
What stops faculty teaching to the dashboard?
Nothing technical, and it is a fair concern. The mitigation is that mastery is computed from concept-tagged questions across many assessments rather than from one weighted exam, so narrowing teaching to the metric means teaching the concepts. Where a metric could be gamed cheaply, we would rather not display it.
Look at a real cohort, not a sample one
Bring one programme and one semester. We will walk through what the dashboard would show for it and what a department would do differently on the strength of it.
Keep reading
-
Higher Education
The full picture for colleges and universities.
-
At-risk students
Turning these signals into an early flag and an intervention.
-
AI quiz generator
The concept tagging that makes a result readable in the first place.
-
AI
Asking the data a question in plain language.
-
For leadership
Coverage, faculty workload and institution-wide trend.