The Purpose of Data Cleaning and Preparation

Record the export date, source system, file name, and storage location. Raw data should not be overwritten.

Cleaning checks and derivations should be written in scripts or documented specifications rather than reconstructed from memory.

Query outputs should identify records requiring review and should include enough information for follow-up without exposing unnecessary identifiers.

Outputs may include quality summaries, query listings, cleaned datasets, analysis datasets, logs, and reports. Each output should have a purpose.