Data Steward Track

Complete this track if you organize shared datasets, checkpoints, logs, experimental results, or handover archives.

Prerequisites#

• IT and Research Safety Core
• The project's information owner and supervisor are identified

Competencies#

You can:

• classify project information and document approved services;
• distinguish source data, working data, temporary data, and final records;
• choose NAS, Euler project/work, scratch, or local storage intentionally;
• preserve provenance, licenses, consent, and access restrictions;
• set collaborative permissions without making data world-writable;
• prevent Git repositories from becoming shared data stores;
• plan backup, retention, cleanup, and project handover;
• restore from snapshots or contact the correct support owner.

Required Learning#

1. Data and AI policy
2. NAS guide
3. Euler storage, when relevant
4. Data placement lab
5. Project handover lab

Evidence#

• A completed data-location map.
• A short README describing data owner, source, classification, access, environment, retention, and recovery.
• A safe shared-folder design for datasets, checkpoints, logs, and results.
• A .gitignore that excludes local data and generated artifacts.

Understand Before Accepting AI Output#

• I verified paths before permission or deletion commands.
• I did not use recursive chmod 777.
• I know which locations are backed up and which are temporary.
• I know who approves access and data disposal.

Pass Criteria#

The data map is approved by the information owner or supervisor and another researcher can understand where authoritative and temporary copies belong.