Cohort analytics

Each cohort has an analytics dashboard that shows how the group is doing and
flags students who may need help — without exposing private detail.

What the dashboard shows

• Overview — active students, enrollment, and engagement at a glance.
• Enrollment trend — how the roster has grown over recent days.
• Attendance heatmap — attendance across recent weeks, so a slipping pattern

is easy to see.

• At-risk students — a privacy-safe list of who needs attention.
• AI engagement — how much the cohort is using the AI tutor.

The dashboard reads from a daily snapshot, so numbers are stable through the day
rather than shifting on every refresh.

The at-risk list is privacy-safe

The at-risk list is deliberately PII-free. Each entry shows only:

• the student's initials (never the full name),
• a severity,
• the reasons they were flagged (low attendance, low grade, inactivity, low

assignment submission, low video completion),

• their grade and days since last activity.

Because there are no names or ids on this list, its actions are cohort-level:
you can open the full roster, or message the whole cohort with a prefilled
template (see Announcements). To act on a specific student, use
the roster.

How "at-risk" is decided

A student is flagged when a signal crosses your cohort's thresholds — for example
attendance below the cohort minimum (75% by default) or a grade below the pass
threshold (60% by default). Both thresholds are overridable per cohort, so the
flag matches the bar your program actually holds students to. A brand-new student
with no data yet is never flagged.

Using it well

Check the at-risk list weekly. A single low signal early is far cheaper to fix
than a disengaged student at the end of term.