Project Step: Create a Data Audit Note

Unit ID: ML-M02-U07 Estimated active time: 35-45 minutes

Your Module 2 evidence

Now create a short data audit note for learner_progress_audit_v1.csv.

The note should help a reviewer decide whether the dataset is ready for a first modelling experiment.

Data audit structure

Use this structure:

Strong audit example

A strong audit does not say:

The data looks fine.

It says:

One row represents one fictional learner's first Module 1 attempt. The day-3 activity columns are candidate features. final_quiz_score, completion_recorded_day, and mentor_message_sent_day4 must be excluded because they are known after the prediction time or after the action. access_bandwidth_band and preferred_device need proxy-risk review. The dataset is suitable for a learning exercise, but it is too small and synthetic for real conclusions.

Decision

Choose one:

• Continue: ready for a simple learning experiment.
• Revise: useful, but a data issue must be fixed first.
• Stop: not suitable or not safe for this modelling purpose.

For this course dataset, the best decision is usually:

Continue for a learning exercise, not for real-world use.

Takeaway

A data audit is the bridge between problem framing and modelling. It keeps the next model honest.