Add D4D datasheet: AI-READI Flagship Dataset

[d4dassistant]

Summary

Created new D4D datasheet for AI-READI Flagship Dataset of Type 2 Diabetes based on comprehensive documentation from:

• https://aireadi.org/ (project overview)
• https://docs.aireadi.org (technical documentation)
• https://fairhub.io (distribution platform)
• https://github.com/AI-READI (code repositories and tools)

Files Added

• data/extracted_by_column/AI_READI/ai_readi_flagship_d4d.yaml - D4D YAML datasheet
• src/html/output/D4D_-_AI-READI_FAIRHub_v3_human_readable.html - HTML preview (updated)

Key Metadata Extracted

Dataset Overview

• Dataset ID: ai-readi-flagship-v2.0.0
• Dataset Name: Flagship Dataset of Type 2 Diabetes from the AI-READI Project
• Publisher: AI-READI Consortium
• Version: 2.0.0 (released July 31, 2024)
• DOI: 10.5281/zenodo.10642459

Purpose

Create and share a flagship ethically-sourced dataset of type 2 diabetes optimized for AI/ML analysis. The dataset illuminates salutogenic pathways from diabetes toward health recovery through computational methods and emphasizes ethical data sourcing and equity.

Composition

• Participants: 1,067 participants (target: 4,000)
• Data Modalities: 15+ types including vitals, ECG, ophthalmological measurements, retinal imaging, wearables
• Dataset Size: 2.01 TB with 165,051 files
• Formats: DICOM, CSV, Markdown
• Categories: 4 diabetes severity groups (no diabetes, prediabetes, oral/non-insulin controlled, insulin-controlled)

Collection

• Sites: UAB, UCSD, University of Washington
• Protocols: Clinical Dataset Structure (CDS) specification
• Devices: Dexcom, Topcon Healthcare, Heidelberg Engineering, iCare, Optomed

Distribution

• Platform: FAIRhub (https://fairhub.io)
• Access: Controlled access with use restrictions
• License: AI-READI custom license v1.0
• Restrictions: Limited to type 2 diabetes-related research
• Standards: FAIR principles, CDS, Bridge2AI standards

Ethics & Privacy

• IRB Approval: All three collection sites
• De-identification: PHI removed, HIPAA Safe Harbor approach
• Community Engagement: Stakeholder input and American Indian engagement initiatives
• Ethical Position: Ethically-sourced with emphasis on equity

Funding

• Agency: National Institutes of Health (NIH)
• Program: Bridge2AI
• Award: 1OT2OD032644

References

• Owsley et al. (2025). BMJ Open. DOI: 10.1136/bmjopen-2024-097449
• AI-READI Consortium (2024). Nature Metabolism. DOI: 10.1038/s42255-024-01165-x

How to Review

1. View HTML preview: Open src/html/output/D4D_-_AI-READI_FAIRHub_v3_human_readable.html in a browser for human-readable format
2. Check YAML: Review data/extracted_by_column/AI_READI/ai_readi_flagship_d4d.yaml for completeness and accuracy
3. Verify sources: Compare against original documentation URLs listed above
4. Check metadata sections: Ensure key information is captured accurately

Notes

• This datasheet follows the same format as other extracted datasheets in the repository (VOICE, CM4AI, CHORUS)
• Uses free-form YAML structure compatible with the HTML renderer
• Comprehensive coverage of all major D4D sections: motivation, composition, collection, preprocessing, distribution, ethics, funding, maintenance

Related to: #58

---

🤖 Generated with D4D Assistant