Live local demonstration dataset
Student Performance Dashboard
Monitor predictions, attendance, academic risk, and program-level patterns. The complete demonstration dataset is loaded automatically and remains editable in the browser.
Performance trend
Illustrative period trend ending with the current dataset average.
Risk distribution
Current risk classification calculated from scores and attendance.
Priority monitoring
Students ordered by risk and predicted score.
Program summary
Average prediction grouped by academic program.
Complete CRUD workspace
Student Explorer
Create, read, update, duplicate, delete, search, filter, sort, import, export, and recalculate student prediction records. Data is visible immediately and stored locally.
Create student record
Add or edit a student. The prediction and risk level are recalculated automatically.
Dataset operations
Manage the complete local dataset without relying on an unavailable API.
The dataset starts with 12 complete records. Empty or invalid old storage is ignored, preventing the project from opening with “0 students” unless the user intentionally deletes every record.
| ID / semester | Student / program | Attendance | Study | Previous | Assignments | Quiz | Prediction | Risk | Operations |
|---|
Interactive local model
Prediction Model Lab
Adjust academic indicators and see the predicted score and risk update instantly. The model is an explainable demonstration formula, not a production admission or grading system.
Scenario inputs
Predicted score0
Risk class—
Demo confidence0%
Explainable formula
prediction = previous × 0.36 + assignments × 0.22 + quiz × 0.16 + attendance × 0.14 + engagement × 0.07 + normalized study time × 0.05
High risk: predicted score below 60 or attendance below 70%. Medium risk: predicted score below 78 or attendance below 85%. Otherwise the profile is classified as low risk.
Input validation
All numeric fields are clamped to valid ranges and invalid values resolve safely instead of producing NaN.
Explainability
Every output can be traced to visible weighted inputs, avoiding an opaque or unverifiable prediction.
Responsible use
Predictions are decision-support signals only and require human academic review.
Group-level quality review
Fairness Audit
Compare average predictions and high-risk rates across dataset groups. Small samples are displayed explicitly and should not be treated as statistically conclusive.
| Group | Records | Avg prediction | High-risk rate | Gap vs overall | Interpretation |
|---|
Responsible interpretation
Fairness analysis should consider sample size, measurement quality, missing context, and whether features create indirect disadvantage. A numerical gap is a prompt for investigation, not proof of discrimination.
Actionable support planning
Student Recommendations
Interventions are generated from attendance, study time, assignment, quiz, and engagement patterns for all medium- and high-risk records.
Complete printable summary
Performance Report
Review dataset size, predicted outcomes, attendance, high-risk records, and intervention priorities in one printable report.