Reproducible Scripted Cleaning

Reproducible scripted cleaning means writing your data preparation as code — a script that reads the raw file and produces the clean one — instead of editing cells by hand. Anyone can re-run the script and get exactly the same result.

This is the single most important habit in data wrangling. Manual edits are invisible and unrepeatable: nobody can tell what you changed or redo it when the data updates. A script is a written record, can be reviewed and version-controlled, and keeps the original raw data untouched so mistakes are always recoverable.