Data Cleaning with Claude

Data work breaks down when inputs are inconsistent, incomplete, or hard to trust. In this course, data analyst Chris French shows you how to upload a real CSV to Claude and clean it end to end through structured conversation—building the same data-quality instincts you use in production analytics. Learn how to prompt Claude to profile a dataset, diagnose issues that affect downstream metrics, resolve missing values and duplicates, standardize messy fields, and validate results with targeted checks. Plus, document decisions and turn your prompts into a repeatable workflow you can reuse across stakeholder exports, ad hoc analyses, and pipeline handoffs.

Learning objectives
Prompt Claude to audit and profile a CSV dataset, surfacing structure, data types, null patterns, and inconsistencies.
Diagnose common data-quality issues—including duplicates, missing values, formatting inconsistencies, and outliers—and prioritize what to fix.
Evaluate and apply appropriate cleaning strategies for duplicates and missing data, based on analytic intent and trade-offs.
Standardize and transform messy fields—including text, categorical, date, and numeric data—using reusable prompting patterns.
Validate cleaned datasets through post-cleaning audits and targeted checks to ensure they are analysis-ready.
Create a repeatable, documented cleaning workflow, including a clear log of decisions for transparency and reuse.