Navigating Data Acquisition and Data Quality Validation of Large Databases: Practical Lessons
Bibliographic record
Abstract
ObjectiveHave you ever been frustrated when working to acquire data? Have you experienced the trials and tribulations of readying data? ‘Best laid plans often go awry’. ‘A problem shared is a problem halved’ and we’d like to share our experiences and lessons learned with you. We are a Canadian provincial organization with over 20 years experience working with large comprehensive data currently in the midst of validation work on newly acquired data. This behind the scenes and sometimes forgotten work can take considerable time, but it is crucial and integral for research. ApproachData acquisition is complex, particularly when adding new data to existing data files. Many validation steps are vital to ensuring that data used in research are the highest quality possible and limitations are understood. The process began with preparing the agreement and list of new variables to add in 2018 for 4 databases (hospital, drug, physician, emergency), to signatures in 2023, and it is ongoing. While we have a plan to do this work, it requires flexibility and constant updates based on new information learned. ConclusionInvestigating errors, deciding level of acceptance of errors, managing relationships, and communication are the most critical aspects of the work. Contingency plans are important when it takes longer than expected. Data are never perfect, but a thoughtful approach to validation and documentation of limitations and decisions support using the data appropriately. Also, a kind demeanor, an open mind, and a sense of humour can go a long way!
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 | 0.223 | 0.389 |
| Meta-epidemiology (narrow) | 0.003 | 0.002 |
| Meta-epidemiology (broad) | 0.002 | 0.003 |
| Bibliometrics | 0.006 | 0.006 |
| Science and technology studies | 0.009 | 0.017 |
| Scholarly communication | 0.023 | 0.040 |
| Open science | 0.011 | 0.017 |
| Research integrity | 0.011 | 0.026 |
| Insufficient payload (model declined to judge) | 0.011 | 0.006 |
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; the direct Gemma label and the distilled Codex classifier agree on what is shown here.
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".