The problem
Dashboards surface risk scores without the records behind them, silently drop conflicting data, and stop at the alert. Nobody owns the next step or the review.
Workflow
- 01See dated attendance and learning trends with thresholds and denominators.
- 02Open work samples, a teacher note, and a list of missing sources.
- 03Confirm or leave unresolved a conflicting record and watch the flag update.
- 04Separate observations from hypotheses.
- 05Choose an intervention with an owner and review date, then review a simulated follow-up.
Key design tradeoff
Simple thresholds over predictive scores
Fixed, visible thresholds are less sophisticated than a risk model, but a leader can check every number by hand and see how one corrected record changes the picture.
What you can try
- Compare Amara (attendance and learning) with Theo (attendance only).
- Confirm Theo's conflicting record as present and watch the flag clear.
- Download the intervention and review plan.
Current limitations
- Two fictional students and six weeks of data.
- Flags are transparent rules, not validated predictions.
- The follow-up is simulated. There are no real integrations.
How real impact would be measured
- Share of flags with a named owner and review date within a week.
- Share of conflicting records resolved by a person rather than dropped.
- Attendance and work completion at review compared with similar students, with small-n caution.