On this page
Choose a bounded fictional task
Use a synthetic business-survey or case-processing problem with a clear population and quantity of interest. Avoid employer or client records in a public portfolio. A small invented dataset can show the reasoning without implying access to proprietary research.
Show the frame and allowed selections
Explain one candidate design and why its groups or ordering matter. For a small illustrative population, list all allowed samples or show a reproducible selection specification. Describe missing units or response risks rather than claiming automatic representativeness.
Include a calculation a reviewer can inspect
| Artifact | Evidence of skill |
|---|---|
| Question and frame definition | Scope and unit discipline |
| Selection plan | Design and probability reasoning |
| Weighted estimate or exact distribution | Auditable arithmetic |
| Limitations and alternative design | Judgment about the conclusion |
Keep claims about credentials separate
Label the dataset fictional and the analysis educational. Do not imply that completing the project grants a professional designation or proves a commercial outcome. Include a concise explanation of what would need validation before adapting the design to real records.