WORKBENCH NOTE
Use this to keep a time-boxed assignment focused on a clear answer and reliable evidence. Complete the core work before adding sophistication.
Before starting
- Clarify the ask, audience, deliverables, time constraint, and assumptions.
- Write a short plan that maps the core question to the minimum analysis needed to answer it.
Must have
- Valid analysis with essential data-quality checks.
- A baseline when modeling is involved and an appropriate validation approach.
- A direct answer to the core question and a clear evidence-based recommendation.
Should have
- Useful visuals, explicit limitations, a concise README, and a reproducible flow.
- Labels and narrative that help a reviewer understand the decision without reading every line of code.
Nice to have
- An extra model, deeper segmentation, or a dashboard only after the core work is complete.
- Do not trade a clear answer for a longer list of optional additions.
Before submission
- Rerun the work, remove dead experiments, and check labels, conclusions, and assumptions.
- Check the README and state unfinished ideas honestly.
- Make sure each recommendation follows from the analysis rather than from an unsupported claim.