Data quality before models on the AI Learning track. Duplicates, wrong types, leaked columns, and timezone mix-ups dominate model quality. A clean smaller set beats a huge dirty one. Check this before architecture debates.
This lesson assumes you already worked through Error analysis.
The idea in practice
Profile nulls, ranges, and duplicate keys. Decide a unit. Drop or repair rows with a written rule.
A concrete check
goal = {
'track': 'AI Learning',
'lesson': 'Data quality before models',
}
checks = [
'input available at decision time',
'score matches the real decision',
'failure case written down',
]
print(goal['lesson'])
for item in checks:
print('-', item)
Run the sketch locally if you have Python. The printout is a reminder of the checks, not a trained model. Replace the strings with the real inputs from your own example before you treat it as a design.
What usually goes wrong
Imputing missing targets and training as if they were real. When this happens, stop adding parameters or tools. Fix the check, the data, or the permission, then run the same example again.
What to write down
- The input you are allowed to use at decision time.
- The output and the score or pass rule.
- One failure you will test on purpose.
- What you will not claim the system can do.
Practice
Write three data checks you would fail a training job on.
Self-check
- Say Data quality before models in one sentence that mentions an input and an output.
- Name the failure mode in this lesson and the check that would catch it.
Done when: you can explain this lesson without the page open, and you have a written failure case.