Data Analyst Role in EdTech: What the Job Really Looks Like

-- Weekly active learners by course
SELECT course_id,
       strftime('%Y-%W', event_time) AS year_week,
       COUNT(DISTINCT learner_id)     AS weekly_active_learners
FROM lesson_events
WHERE event_type IN ('lesson_start', 'video_progress', 'quiz_attempt')
  AND event_time >= '2026-09-01'
GROUP BY course_id, year_week
ORDER BY course_id, year_week;

import pandas as pd

ev = pd.DataFrame({
    "learner": ["a", "a", "a", "b", "b", "c"],
    "week_since_enrol": [0, 1, 3, 0, 1, 0],
})
size = ev.loc[ev["week_since_enrol"] == 0, "learner"].nunique()
active_share = ev.groupby("week_since_enrol")["learner"].nunique() / size
print((active_share * 100).round(0))
# Shows when learners typically stop engaging.

-- Share of enrolled learners reaching each module
WITH reached AS (
  SELECT module_no, COUNT(DISTINCT learner_id) AS learners
  FROM module_progress
  WHERE course_id = 'DA-101'
  GROUP BY module_no
), base AS (
  SELECT COUNT(DISTINCT learner_id) AS enrolled
  FROM enrolments WHERE course_id = 'DA-101'
)
SELECT r.module_no, r.learners,
       ROUND(100.0 * r.learners / b.enrolled, 1) AS pct_reached
FROM reached r CROSS JOIN base b
ORDER BY r.module_no;

import pandas as pd

answers = pd.DataFrame({
    "question": ["Q1", "Q1", "Q1", "Q2", "Q2", "Q2", "Q3", "Q3", "Q3"],
    "correct":  [1, 1, 0, 0, 0, 1, 1, 1, 1],
})
difficulty = (answers.groupby("question")["correct"].mean()
                     .rename("pct_correct").sort_values())
print((difficulty * 100).round(0))
# Lowest values flag questions to review for clarity or teaching gaps.

import pandas as pd

funnel = pd.DataFrame({
    "step": ["Visited page", "Signed up", "Attended demo", "Paid"],
    "users": [20000, 3200, 1400, 520],
})
funnel["from_prev_%"] = (funnel["users"] / funnel["users"].shift(1) * 100).round(1)
funnel["from_start_%"] = (funnel["users"] / funnel["users"].iloc[0] * 100).round(2)
print(funnel)
# Illustrative numbers for practice only.

-- Renewal rate by how much of the course learners completed
SELECT CASE
         WHEN completion_pct >= 75 THEN '75-100%'
         WHEN completion_pct >= 40 THEN '40-74%'
         ELSE 'under 40%'
       END AS completion_band,
       COUNT(*) AS learners,
       ROUND(100.0 * AVG(CASE WHEN renewed THEN 1 ELSE 0 END), 1) AS renewal_pct
FROM learner_period_summary
WHERE period_end BETWEEN '2026-07-01' AND '2026-09-30'
GROUP BY completion_band
ORDER BY completion_band;

learners(learner_id, signup_date, channel, city)
enrolments(learner_id, course_id, batch_id, enrol_date)
lesson_events(learner_id, course_id, lesson_id, event_type, event_time)
module_progress(learner_id, course_id, module_no, completed_at)
quiz_answers(learner_id, quiz_id, question_id, correct, attempt_no)
learner_period_summary(learner_id, period_end, completion_pct, renewed)

EdTech learning dashboard

- Weekly active learners (trend, by course)
- Module reach chart for one course
- Hardest 10 questions (lowest % correct)
- Funnel: visit -> sign-up -> demo -> paid
- Renewal rate by completion band

Filters: course, batch, channel, date range

Finding: Learners who complete fewer than 40% of modules renew
much less often than those who complete more.

Evidence: Renewal by completion band, last quarter.

Recommendation: Mentor check-ins for learners below 40%
completion by week four; measure the effect next quarter.

Caveat: Correlation; engaged learners may differ in other ways.