Activity: Create a Data Audit

Activity ID: ML-M02-A01 Estimated active time: 40-55 minutes

Goal

Create a data audit for the supplied synthetic learner-progress dataset.

Inputs

• data/learner_progress_audit_v1.csv
• data/data-dictionary.md
• data/schema.json
• data/provenance.md

Instructions

1. Read the data dictionary and note which columns are recorded after the outcome.
2. For each column, write when it becomes available relative to the prediction moment.
3. Mark every post-outcome column as not a feature candidate, and say why.
4. Identify the repeated unit: does one row equal one learner, or do learners appear several times?
5. Sketch how the rows were generated — who is in this data and who is missing.
6. Write two limitations that follow from the generating process, not from the model.

Submission table

Quality checklist

• [ ] Every column is classified as available-before or recorded-after.
• [ ] final_quiz_score and completion_recorded_day are excluded, with the reason stated.
• [ ] The repeated unit is identified.
• [ ] The note says who is missing from the data, not only who is in it.
• [ ] Limitations follow from how the data was produced.

Safety note

Use only the supplied synthetic data. Do not add or upload real learner data.