Open science
Research that cannot be checked is not much use to anyone. These are the standards REACH work is held to, from where a dataset is stored through to how a release is licensed.
Data principles
A code repository is not a data warehouse. Where a dataset lives depends on what it is and what we are legally allowed to do with it.
| Material | Storage |
|---|---|
| Analysis code | Public code repository |
| Protocols and documentation | Public code repository |
| Small openly licensed derived files | Public code repository |
| Raw government data | Left with the original custodian, retrieved by a download script |
| Large public releases | Zenodo, OSF, Dataverse, or an institutional repository |
| Restricted data | Approved institutional storage only |
| Partner confidential files | Approved restricted drive |
| Credentials | A password manager or repository secrets, never in a file |
When we cannot redistribute a dataset, we publish the download script, the source URL, a checksum, and the instructions to rebuild it. That keeps the analysis reproducible without republishing data we have no right to move.
Credit and authorship
Authorship follows ICMJE criteria and documented contribution, and nothing appears publicly until it is released. Both are set out in full under how we govern the work.
Reuse and licensing
Each release carries an explicit license, a citation file, and a version. Where a release is archived, the DOI and version are recorded in the project README so a reader can cite the exact state of the work they used.
Our public code repositories will be linked here once the first release is published. Until then there is nothing public to point you at, and we would rather say so than link an empty page.