Data quality
Data quality is learned and done in 8 activities, made of 136 knobits.
Learning path
- Set the data quality strategy and framework · 12 knobits
- Select the critical data elements · 10 knobits
- Document data rules in SBVR · 23 knobits · builds on Select the critical data elements
- Derive data quality rules · 30 knobits · builds on Document data rules in SBVR
- Profile and assess data quality · 10 knobits · builds on Select the critical data elements, Derive data quality rules
- Analyse the root causes · 9 knobits · builds on Profile and assess data quality
- Manage data quality issues · 9 knobits · builds on Analyse the root causes
- Set data quality service levels and the scorecard · 10 knobits · builds on Select the critical data elements, Profile and assess data quality
Foundation knowledge
23 knobits used across the activities rather than in one of them.
- Q-901 Explains the business drivers of data quality management
- Q-902 Explains the goals and principles of data quality management
- Q-903 Explains data quality as fitness for purpose
- Q-904 Explains the published frameworks of data quality dimensions
- Q-905 Explains the role of metadata in data quality
- Q-906 Explains the parts of the ISO 8000 data quality standard
- Q-907 Explains the types of data quality business rules
- Q-908 Explains the common causes of data quality issues
- Q-909 Explains data cleansing, enhancement, parsing and standardisation
- Q-910 Explains the categories of data quality tools
- Q-911 Explains the kinds of preventive and corrective data quality actions
- Q-912 Explains reusable quality check and audit code modules
- Q-913 Explains the characteristics of effective data quality metrics
- Q-914 Explains statistical process control for data quality
- Q-915 Explains the readiness and risk assessment for a data quality programme
- Q-916 Explains the organisational and cultural change a data quality programme needs
- Q-917 Explains the content of a data quality policy
- Q-918 Explains the metrics of a data quality programme
- Q-1801 Defines the concepts of sample and population
- Q-1802 Explains the concept of the arithmetic mean
- Q-1803 Explains the meaning of the set operations union and intersection
- Q-1804 Explains the concept of a percentage
- Q-1806 Explains the meaning of negation, conjunction and disjunction
Which AI model is best at data quality?
Best result for data quality: gemini-4-argon (89% test score). Best value: claude-haiku-5-5 (89%, within 5 points of the best at $0.006 per run instead of $0.021).
| Model | 1 | 2 | 3 | 4 | 5 | 6 | 7 | 8 | All |
|---|---|---|---|---|---|---|---|---|---|
| gemini-4-argon | 63% | 78% | 89% | 82% | 77% | 88% | 84% | 81% | 80% |
| gemini-3.1-pro-preview* | – | – | 86% | 86% | – | – | – | – | 86% |
| gpt-6-astra* | – | – | 94% | 98% | – | – | – | – | 96% |
| claude-fable-5-1* | – | – | 98% | 95% | – | – | – | – | 96% |
| gemini-3.8-flash* | – | – | 93% | 94% | – | – | – | – | 94% |
| gpt-6.1-sol* | – | – | 94% | 98% | – | – | – | – | 96% |
| claude-sonnet-5-5* | – | – | 100% | 92% | – | – | – | – | 96% |
| gemini-3.5-flash-lite | 33% | 13% | 43% | 58% | 52% | 78% | 55% | 53% | 48% |
| gpt-6-luna | 48% | 50% | 78% | 74% | 68% | 88% | 81% | 75% | 70% |
| claude-haiku-5-5 | 48% | 75% | 85% | 87% | 66% | 94% | 85% | 76% | 77% |
* From the earlier benchmark of 2026-10: 10 data-modelling test tasks, 3 runs per task, graded by claude-opus-5-5. These models have not taken the current 123 tasks yet; the two methods differ, so compare their figures with care. They are left out of the “most models” counts and the best-model picks.
Expected share of each activity's knobits the model masters; columns follow the learning path above. Method and caveats.