School analytics
Who this is for: directors · Where: Analytics in the sidebar.
What this page is for
The school at a glance: how pupils are doing, what staff are producing, how much AI is being spent, and how many parents you are actually reaching.
The header carries three quick badges — paid pupils, pupils awaiting payment, and the school average — plus a system-health chip. Below it, four cards give the current headcount: teachers, pupils, official classes, and staffing coverage (the share of teaching assignments that are actually staffed). A further row of counters covers homework awaiting review, homework submissions, exam sessions, catalogue grade entries, teaching assignments, and recent audit events.
The academic year filter
Most of the page is scoped to one academic year, chosen at the top. It defaults to the current one.
The academic year flips on 1 June, so from June onward the new year is selected by default and will look sparse until marks start arriving. Switch to the previous year to see the completed one.
Grades and attendance follow the selected year. AI usage is the last 30 days, and headcounts are current. Mixing those on one screen is deliberate — "how did last year go" and "what are we spending right now" are different questions.
Academic
- Grade distribution across the school
- Performance by subject
- Grade trend — the monthly school average
- Class ranking
- Enrolment by grade
- At-risk distribution and a per-class breakdown, mirroring Pupils at risk
AI usage
The last 30 days: total tokens, estimated cost, a breakdown by feature (exams, homework, lesson plans, presentations, videos, audio, concept maps, images, games), and a daily series.
This is where "why did we run out of tokens" gets answered. One feature usually dominates.
This section sits directly below the academic charts (enrollment by grade, attendance) — a reminder that the two are scoped differently. The charts above follow the academic year selector; the ones here are always the last 30 days, no matter which year is selected.
Parent reach
How many pupils have a linked parent, expressed as a percentage, plus how many parents have connected Telegram and how many announcements have gone out.
A grade alert nobody receives is not an alert. If the linked percentage is low, that is usually the highest-value thing on this page to act on — before any academic metric.
Revenue
Your price per pupil, how many pupils are covered, and the resulting monthly figure in MDL.
This is a straight multiplication of the configured price by the number of covered pupils. It is not an accounting record and does not know about refunds or partial periods.
Content production
How much staff have created: lesson plans, exams, homework, presentations, videos, audio, concept maps, games, materials.
Rules and limits
- Directors only.
- Grades and attendance follow the selected year; AI usage is a rolling 30 days; headcounts are current.
- The year list is built from real data, plus the current year, so a fresh year is always selectable.
- Revenue is price × covered pupils — an indicator, not accounting.
- Aggregates only — no individual pupil is identified here.
Troubleshooting
"Everything is empty." You are probably on the new academic year, which begins on 1 June. Switch to the previous one.
"AI usage does not match the token pool." This chart is 30 days. The pool resets monthly on its own schedule. They will rarely line up exactly.
"Revenue looks too low." It counts covered pupils. Unpaid pupils, or a school trial where nobody has paid yet, both reduce it.
"Class ranking has a class missing." It needs marks in the selected year. A class with no grades yet cannot be ranked.