THE CORE AIM OF VOCABULARY MASTERY · DATA QUALITY VOCABULARY · DEFINE → TEST → DETECT → REPAIR → TRUST
Data quality vocabulary is the language used to describe whether data is fit for its intended use. Terms such as accuracy, completeness, validity, consistency, uniqueness, timeliness, freshness, duplicate and quality rule matter because “bad data” is too vague to diagnose or repair.
The core aim of vocabulary mastery for data quality vocabulary is trust clarity. Learners should be able to state what quality dimension is failing, what rule detects the failure, which records are affected, who owns the correction and what threshold is acceptable for the business purpose.
This page is the Data Quality Vocabulary owner inside the eduKateSG Vocabulary hub. For ownership and policy, use Data Governance Vocabulary. For pipelines, use Data Engineering Vocabulary.
Central proposition: Data quality vocabulary is mastered when “the data is wrong” can be replaced by a specific, measurable quality failure with an owner and repair path.
The 60-Second Data Quality Vocabulary Router
- Accuracy: correct, verified, source-of-truth comparison.
- Completeness: missing, null, required field, coverage.
- Validity: format, range, allowed value, rule.
- Consistency: conflict, mismatch, reconciliation.
- Uniqueness: duplicate, key, deduplication.
- Timeliness: freshness, delay, stale, update frequency.
The Data Quality Vocabulary Architecture
| Dimension | Core terms | Core question |
|---|---|---|
| Accuracy | correct, verified | Does the value reflect reality? |
| Completeness | missing, null, coverage | Is required data present? |
| Validity | format, range, rule | Does the value conform to expectations? |
| Consistency | mismatch, reconciliation | Do sources agree? |
| Uniqueness | duplicate, key | Is the same entity represented more than once? |
| Timeliness | freshness, delay | Is the data current enough for the use case? |
Accuracy and Validity Are Different
A value can be valid without being accurate. A date may match the required format yet still record the wrong day. Validity checks whether the value follows rules; accuracy asks whether it reflects reality.
A Worked Example: Completeness
If customer records require an email address for a campaign, completeness can be measured as the proportion of eligible records with a usable email value. The vocabulary becomes operational when the field, population and threshold are explicit.
A Worked Example: Freshness
Freshness describes how recently data was updated relative to what the use case requires. A daily report may tolerate hours of delay; a fraud-monitoring system may require seconds. Freshness therefore depends on purpose.
Duplicates and Uniqueness
Uniqueness asks whether one real-world entity or event is represented only once where that is expected. Duplicate detection may rely on exact keys or fuzzy matching when identifiers are incomplete or inconsistent.
Data Quality Rules
A useful quality rule states what field or dataset is checked, what condition must hold, what threshold is acceptable and what happens when the rule fails. Examples include “order amount must be non-negative” or “daily sales data must arrive before 8 a.m.”
Quality Issue and Root Cause
A quality issue is the observed defect; the root cause may be a source-system bug, manual entry problem, transformation error or delayed pipeline. Repairing records without addressing the source can cause the same issue to return.
How to Learn Data Quality Vocabulary
- Choose one dataset and profile it.
- Write rules for each quality dimension.
- Measure missing and duplicate rates.
- Compare source and downstream values.
- Track freshness explicitly.
- Assign issue owners.
- Separate symptom repair from root-cause remediation.
Common Data Quality Vocabulary Mistakes
Using quality as one vague score
Repair: name the specific dimension.
Confusing validity with accuracy
Repair: distinguish rule conformance from truth.
Fixing records without fixing the source
Repair: trace recurring issues to root cause.
Ignoring business context
Repair: define acceptable thresholds according to intended use.
Frequently Asked Questions
What is data quality vocabulary?
It is the language used to describe accuracy, completeness, validity, consistency, uniqueness, timeliness and quality controls.
What terms should beginners learn first?
Start with accuracy, completeness, validity, consistency, duplicate, uniqueness, freshness and quality rule.
What is the difference between accuracy and validity?
Validity asks whether a value follows rules; accuracy asks whether it reflects the real-world value correctly.
What is data freshness?
It is how recently data was updated relative to the timeliness required by the use case.
How can I learn data quality vocabulary?
Profile a real dataset and turn vague problems into explicit, measurable quality rules.
Where This Article Fits in the eduKateSG Vocabulary Ecosystem
- Vocabulary Hub — the broad route.
- Data Governance Vocabulary — accountability.
- Data Engineering Vocabulary — pipelines.
- Data Analytics Vocabulary — decision use.
- Business Intelligence Vocabulary — reporting trust.
The Data Quality Vocabulary Standard
Data quality vocabulary reaches its core aim when the learner can name the quality dimension, define the failing rule, quantify the impact and assign a repair path.
That is the standard: quality language precise enough to turn distrust into measurable improvement.