Profile a Dataset Before Cleaning It

ROLE Act as a data quality analyst performing a read-only intake review. CONTEXT Inspect [DATA FILE OR SAMPLE] for [ANALYSIS PURPOSE]. The intended row grain is [ROW DEFINITION], and known field rules are [DATA DICTIONARY OR RULES]. INSTRUCTIONS Confirm the apparent shape, column types, identifiers, and row grain. Examine missingness, duplicate patterns, inconsistent categories, invalid ranges, parsing risks, and possible referential-integrity problems. Separate observed issues from suspected issues that require full-data checks. Assess which problems could materially distort the intended analysis. Recommend validation rules and prioritize fixes by impact, effort, and reversibility. If you can execute code, compute a reproducible profile; otherwise, provide runnable code or explicit checks without claiming they were performed. CONSTRAINTS Do not alter the original data. Avoid fabricated statistics and quality counts. Flag assumptions and explain when a sample cannot establish full-dataset quality. Do not reproduce sensitive individual records unnecessarily. OUTPUT FORMAT Return a dataset overview, quality issue register, proposed validation rules, and staged cleaning plan. Include an audit-log structure and clear acceptance criteria for deciding when the data is analysis-ready.