A data quality dimension is a specific, measurable characteristic of data that can be evaluated to determine whether the data is fit for its intended use — representing one distinct aspect of quality such as completeness, accuracy, validity, or uniqueness.
No single metric can capture all aspects of data quality. A dataset might be 99% complete but contain many inaccurate values. It might be perfectly formatted (valid) but full of duplicates (not unique). Data quality dimensions provide a structured vocabulary for talking about quality across multiple aspects — allowing organizations to measure each dimension separately and understand exactly where quality is failing.
The Core Data Quality Dimensions
Different frameworks define different sets of dimensions. The most commonly used are:
Completeness: The proportion of required fields that contain non-null, non-empty values. A customer record without an email address is incomplete for email marketing purposes.
Sohovi profiles every column in your dataset for completeness and flags the exact rows where values are missing — free to try.
Accuracy: How closely data values reflect the actual real-world state they're supposed to represent. An address that's formatted correctly but points to the wrong building is valid but inaccurate.
Validity: Whether values conform to defined formats, ranges, or business rules. An email address without "@" is invalid; a discount of 150% is invalid.
Consistency: Whether the same information is represented the same way across records and systems. "Active" and "active" in the same status field is inconsistency.
Uniqueness: Whether entities that should appear only once do appear only once. Duplicate customer records violate uniqueness.
Sohovi automatically finds every duplicate in your dataset — including near-matches — and shows you exactly which rows are affected.
Timeliness: Whether data is sufficiently current for its intended use. A customer's address from 5 years ago may no longer be accurate.
Integrity: Whether relationships between data elements are valid and consistent. An order pointing to a customer ID that doesn't exist violates referential integrity.
Why Dimensions Matter
Dimensions provide the framework for targeted diagnosis. "This data is bad" is not actionable. "This data has 22% null rate on the email field (completeness failure) and 8% duplicate records (uniqueness failure)" is actionable — it tells you exactly what to fix and prioritize.
Sohovi automatically finds every duplicate in your dataset — including near-matches — and shows you exactly which rows are affected.
[IMAGE: A data quality scorecard showing separate scores for each dimension — completeness 94%, validity 99%, uniqueness 92%, consistency 97%]
Sohovi's quality reports score your data across multiple dimensions simultaneously — showing you where each field is strong and where it needs attention. Free to try, no code required.
Frequently Asked Questions
Q: What is a data quality dimension? A data quality dimension is a specific, measurable aspect of data quality — such as completeness, accuracy, or validity — that can be evaluated independently to understand a particular facet of whether data is fit for its intended use.
Q: How many data quality dimensions are there? Different frameworks define different numbers. DAMA's framework uses 6 primary dimensions (completeness, validity, consistency, integrity, timeliness, accuracy). Other frameworks use 10 or more. The exact number matters less than having enough dimensions to cover the quality characteristics relevant to your use cases.
Q: What is the most important data quality dimension? It depends on the use case. For email marketing, completeness (email field) and validity (email format) are most important. For financial reporting, accuracy is paramount. For deduplication, uniqueness is the key dimension. The most important dimension is the one that most affects your primary use case for the data.
Q: What is the difference between validity and accuracy? Validity checks whether a value conforms to a defined format or rule — an email must contain "@." Accuracy checks whether the value reflects reality — the email must actually belong to this person and be deliverable. A value can be valid (syntactically correct) but inaccurate (the wrong email address).
Q: What is the difference between consistency and conformity? Consistency checks whether the same information is expressed the same way across records and systems. Conformity (or standardization) checks whether values follow a defined format convention. They're related but distinct: consistency is about internal agreement; conformity is about adherence to a defined standard.
Q: Can a dataset score well on some dimensions and poorly on others? Yes — and this is exactly why dimensions are useful. A dataset might be 99% complete (almost no missing values) but have 15% duplicate records (poor uniqueness). A dimension-by-dimension score reveals which aspects are strong and which need attention.
Q: How are data quality dimensions scored? Each dimension is typically scored as a percentage: "91% of records are complete" means completeness = 91%. An overall quality score is usually a weighted average across dimensions, with weights reflecting the relative importance of each dimension for the specific use case.
Q: What is data currency and is it the same as timeliness? Data currency refers to how recently data was collected or verified. Timeliness refers to whether data is sufficiently current for its intended use. They're closely related — currency is the measurement (data was last verified 18 months ago), and timeliness is the judgment (is 18-month-old data current enough for this purpose?).
Q: How do different industries weight data quality dimensions differently? Healthcare prioritizes accuracy (wrong patient data can cause harm) and timeliness (medical records must reflect current state). Financial services prioritizes accuracy and integrity (data must reflect actual transactions). Marketing prioritizes completeness (need contact information) and validity (need deliverable email addresses).
Q: What is a data quality score and how is it calculated? A data quality score is an aggregate metric summarizing quality across dimensions. It's typically calculated as a weighted average of dimension scores, where weights reflect the relative importance of each dimension for the use case. Different tools use different weighting schemes.
Data quality dimensions are the language of precise quality measurement. Without them, quality is vague. With them, you can say exactly where data is failing, by how much, and what the impact is.