Co-creating a Data Asset Inventory for Equity-Oriented Research
Bibliographic record
Abstract
BackgroundData asset inventories (DAI) are invaluable resources providing high-level information about datasets, including name, region, purpose, scope, and contents. An effective DAI improves understanding about what data exists, how they can be used, and where they may be limited. Strategies for integrating principles of inclusion, diversity, equity, and accessibility (IDEA) into DAIs are not widely available, which inhibits efforts to improve existing data infrastructure’s capacity for equity-oriented research. ApproachHealth Data Research Network Canada (HDRN) is a pan-Canadian network that works collaboratively to enable innovative multi-regional research that can improve health and health equity. Central to HDRN’s mandate is strengthening researcher capacity to apply IDEA principles within their research. Within this scope of work, we convened a multi-disciplinary team of computer scientists, data experts, and IDEA specialists to build a DAI of Canadian health and social data using de-stigmatizing language. Through discussion the team identified key data categories and co-created definitions. Data sets were annotated accordingly. ResultsDiscussions highlighted disciplinary differences in understandings of IDEA and its relevance to DAIs. Overcoming these differences required IDEA specialists to be both subject matter experts and advocates. Project delays occurred due to the additional time needed for education. Productive conflict resulted in consensus-building to establish respectful and inclusive terms to describe data. DiscussionEmbedding IDEA within health data infrastructure requires a lens that may not be associated with one’s training. Developing that lens within a project is possible but requires additional time and effort. These components must be factored into project planning.
Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.
How this classification was reachedexpand
Full frame machine prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. The Gemma side is a direct model label for every work in the frame, read from the title-only record. The Codex side is a classifier learned from the 10,348 direct Codex labels and calibrated to design-weighted sample rates; fields without enough sample support carry no Codex call. Candidate is the union of the two sides; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels.
Distilled classifier scores by category (both heads)
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.294 | 0.255 |
| Meta-epidemiology (narrow) | 0.001 | 0.002 |
| Meta-epidemiology (broad) | 0.001 | 0.002 |
| Bibliometrics | 0.013 | 0.011 |
| Science and technology studies | 0.015 | 0.029 |
| Scholarly communication | 0.029 | 0.031 |
| Open science | 0.006 | 0.038 |
| Research integrity | 0.004 | 0.013 |
| Insufficient payload (model declined to judge) | 0.011 | 0.003 |
Machine scores (provisional)
The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.
Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.
score_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from itClassification
machine, unvalidatedMachine predicted; a candidate call from one source (direct Gemma or distilled Codex), not a consensus.
How this classification was reached, model by model and score by score, is at the end of the page under "How this classification was reached".