projecthowto/data

Structuring the /data Directory

German version: README_de.md

project-name/
├── data/
    ├── raw/          # Original, unmodified data
    ├── processed/    # Data derived from raw data
    └── models/       # Model weights and outputs (if applicable)

Guidelines

• raw/:

  • Store all original data exactly as collected.
  • Do not modify these files to preserve data integrity.
  • Include metadata or use clear filenames with dates to indicate when the data was recorded.

• processed/:

  • Place cleaned or transformed data here.
  • Generate these files using scripts.
  • Document processing steps if necessary.

• models/ (if you are training models):

  • Save model weights, checkpoints, and outputs.
  • Organize by experiment or model version if needed.
  • Include metadata about training parameters or results.