Data handling standards

Expand me

Please add to this

Data directory structure

• Keep "external" data separate from generated data
• Keep "temporary" data separate from "permanent" data
• Value-added data should be added to a data repository as soon as possible (openICPSR, Dataverse, Zenodo)

(Almost) never do interactive data processing

• You should never do interactive or manual data processing
• If you did it by hand, it should be possible to code it
• If possible, use data-driven programs.
• If nothing else solves the problem, use code that identifies the particular cell or observation, but condition on the "need-to-edit" value (so future changes don't get randomly overwritten).

Downloaded or otherwise acquired data

• Keep in separate folder (e.g., data/external)
• Folder should have a README.md or SRC.md identifying the source, and the procedure to acquire, if necessary.
• Ideally also add a bibtex file with the data citation (optional here, but not optional in references!)

Citing data

• If using data, cite the data! (References)