Faculty Insights
Aggregate student performance and engagement for your cohort, plus question quality signals. You see aggregates only — every student appears as an anonymized ID (e.g. Student A7F3). Emails and names are never shown on this page.
For per-item psychometrics (difficulty, discrimination, distractor analysis), see Analytics → — this page is about students, not items.
No student activity data yet
This dashboard reads the events data stream (practice activity logged by the student apps). That pipeline is still being wired up. This page will activate automatically once practice events start flowing in — nothing for you to configure.
Aggregate performance
Average score over time
Weekly percent-correct from question_answer events. Weeks with no answers are shown as gaps — never as zeroes.
Score distribution
One bar per student: each student's overall percent-correct across the window. Students with fewer than 10 answers are excluded (denominator says so).
Average score by topic / system
Topics with fewer than 10 answers show as “not enough data” instead of a bar.
Engagement
Active students and questions per active student, by week
“Active” = at least one logged event that week. Questions per active student = question_answer events ÷ active students.
Students falling off
Active in the last 30 days but not in the last 7. Identities are anonymized; reach out through your normal channels.
Question flags
Flags are review signals for faculty judgment — the system never deletes items or changes keys on its own. Items with fewer than 10 responses are excluded. For full item analysis (discrimination, distractor pull, most-missed), open Analytics →.
Students who may benefit from outreach
Not a punishment list. These are anonymized students in the bottom quartile of both activity and performance this window — a gentle check-in may help. Nothing here is shared with students or used for grading.
Methodology & caveats
- Denominators are unique students with at least one practice answer in the window, and every statistic labels its own n. With fewer than 10 students in a cohort, statistics are withheld as “not enough data” rather than shown small.
- Time periods are never combined silently. Every chart labels its window; weekly charts show one point per week, and weeks with no data are gaps, not zeroes.
- Blanks are never zeroes. Missing data stays missing on every chart and table.
- Privacy. Student identities are one-way anonymized IDs derived from the account (e.g. Student A7F3); emails and names never appear here. Practice events carry question IDs, topics, and correctness only.
- Source. Live data comes from the Firestore
eventscollection; until that stream is populated this page also reads the device-local demo event log. The source badge above the charts tells you which one you are looking at. - Question flags are classical-test-theory review signals (p < 0.30 too hard, p > 0.90 too easy). Discrimination-based flags live in Analytics, which pools attempts across exams.