Data Wrangling (also called data munging) is the process of transforming and mapping raw data from various sources into a clean, structured format suitable for analysis, visualization, or machine learning. It encompasses cleaning (handling nulls, duplicates, outliers), reshaping (pivoting, melting, exploding nested structures), enriching (joining, deriving new features), and validating (schema enforcement, quality checks). Unlike simple ETL, wrangling is iterative and exploratory—analysts spend 60–80% of project time on it because real-world data is messy: inconsistent formats, missing values, mixed encodings, and unexpected schema drift. Mastering wrangling means knowing not just which tool (pandas, SQL, Spark, Polars, DuckDB) but which technique to apply when—and understanding the performance trade-offs between in-memory operations, lazy evaluation, and distributed processing.