Back to portfolio

Case study · School systems · Interactive prototype

SignalDesk

A decision room that shows why a student is flagged and what a person decided.

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

  1. 01See dated attendance and learning trends with thresholds and denominators.
  2. 02Open work samples, a teacher note, and a list of missing sources.
  3. 03Confirm or leave unresolved a conflicting record and watch the flag update.
  4. 04Separate observations from hypotheses.
  5. 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.