Overview
ALX Data Science: Learner Progression & Completion Intelligence is an EdTech product analytics case study built around the structure of a guided, cohort-based data science programme. The work focuses on where learner momentum breaks, how assignment strain changes dropout risk, and which learner patterns deserve earlier intervention.
Hero
EdTech product analytics case study focused on retention, progression, completion, and dropout pressure in a sprint-based learning programme.
Intelligence Layer
A cohort programme does not fail in one place. Pressure builds through onboarding, pacing, device constraints, assignment deadlines, and middle-stage fatigue. This case joins those signals so the programme team can act earlier and more precisely.
Problem
The core problem is not low content usage by itself. It is the lack of a joined-up view across weekly lesson progress, sprint pressure, assignment behavior, and final learner status. Without that, teams can see that completion is weak without clearly seeing where the breakdown starts.
Data / Signals
Analyst Objective
Build a decision-ready learner intelligence layer that helps teams:
- track retention by cohort,
- isolate the sprints where dropout clusters hardest,
- measure how engagement and assignment behavior change completion odds,
- and identify where learner support should intervene first.
Stakeholders
- Product teams improving learner experience and platform support.
- Programme teams managing cohort pacing and intervention timing.
- Academic leads reviewing sprint load and assignment design.
- Leadership teams tracking completion and learner risk across intakes.
Key Questions
- Where do cohorts lose the most learners over time?
- Which sprint carries the heaviest single drop-off?
- How strongly does steady weekly engagement improve completion?
- When late work rises, how much more vulnerable does the learner become?
- Which device patterns need different support or design?
KPI Framework
- Cohort health: retention curves, completion rate, active rate.
- Progression: weekly completion movement, sprint-level drop-off.
- Engagement: time on platform, lesson completion rate, engagement bands.
- Assignment health: submission rate, lateness, score, dropout gap.
- Learner risk: device pattern, employed-learner pressure, dropout reason mix.
Insight
- The modeled programme covers 538 learners across 10 cohorts, with a 46.7% dropout rate and a 46.1% completion rate.
- Sprint 4 carries the heaviest single drop cluster, but one third of all dropouts still happen by week 6.
- Mobile-only learners complete 16.3 percentage points less often than desktop learners in this run.
- High lateness lifts dropout by 27.2 percentage points versus low-lateness learners.
- 57.8% of week 6 to 10 dropouts come from employed learners, which points to middle-stage schedule pressure.
Implication
- Onboarding still needs structured support, even though Sprint 4 is the sharpest single spike.
- Late work should be treated as an early intervention signal, not only an academic admin issue.
- Mobile learning needs lighter and clearer support for assignment-heavy weeks.
- Employed learners appear to need more flexible pacing or better week 6 to 10 guardrails.
Closing
Deliverables
- Synthetic learner, cohort, progress, assignment, and dropout datasets.
- Pandas-based analysis for retention, sprint pressure, assignment health, and completion risk.
- Exported charts and documented notebook for stakeholder storytelling.
- Hosted walkthrough report plus readable GitHub project documentation.
Outcome
The case demonstrates how EdTech analytics can move beyond activity reporting into a clearer learner-intelligence view that is more useful for retention and programme decisions.
Key Visuals
Retention does not move evenly by cohort, but the broad pattern still shows an early decline followed by a later pressure point.
The sharpest single drop cluster lands in Sprint 4, while the first two sprints still carry heavy early loss.
Steady weekly engagement changes completion odds materially in this model.
Late submissions behave like an early warning signal, not a minor grading detail.
Mobile-only learners carry the weakest completion outcome in this run.
Recommendations
- Strengthen onboarding support in the first two sprints.
- Escalate late work into earlier learner intervention.
- Review Sprint 4 workload and checkpoint design.
- Improve mobile support for assignment-heavy phases.
- Add targeted pacing support for employed learners during weeks 6 to 10.
Embedded Project
Link
- Hosted walkthrough: ALX Product Intelligence
- GitHub repo: alx-product-intelligence