Dimensions Are Connected
The six standard data quality dimensions (completeness, accuracy, consistency, timeliness, validity, uniqueness) don't operate independently. A problem in one dimension can cause or mask problems in others. Understanding these dependencies makes root cause analysis faster and remediation more effective.
Completeness and Accuracy
Incomplete data masks inaccuracy: If 30% of records are missing a value, you can only assess accuracy for the 70% where values are present. Improving completeness (by filling in missing values) may reveal accuracy problems that were previously hidden.
Imputing missing values can introduce inaccuracy: When you fill in missing values (with averages, educated guesses, or enrichment data), you may introduce inaccurate values into previously complete records. Completeness improvement can cause accuracy regression.
Sohovi profiles every column in your dataset for completeness and flags the exact rows where values are missing — free to try.
Accuracy and Consistency
Accurate data that's inconsistently formatted is inconsistent: The same correct value stored in two different formats ("+1 (555) 123-4567" vs. "5551234567") is accurate in both cases but inconsistent. Accuracy and consistency are satisfied simultaneously only when values are both correct AND formatted identically.
Cross-system inconsistency often reveals accuracy problems: When two systems disagree on the value of a field, at least one of them is inaccurate. Consistency checks are often how accuracy problems first surface.
Validity and Accuracy
Valid data is not necessarily accurate: A phone number can pass format validation (10 digits, US number) and still be the wrong phone number for this person. Validity checks are cheaper than accuracy checks — use them first, then verify accuracy for values that pass validity.
Sohovi lets you set up validation rules for any column and instantly see which rows fall outside them — no code or SQL required.
Invalid data is never accurate: A date of 2026-02-30 is both invalid and inaccurate. Invalid data always has an accuracy problem. Validity is a necessary (but not sufficient) condition for accuracy.
Uniqueness and Accuracy
Duplicate records have at least one inaccurate record: When two records represent the same entity, they often disagree on at least one field. Resolving duplicates requires determining which version is accurate.
Uniqueness problems cause accuracy problems in analytics: A customer who appears twice will have their metrics double-counted. Uniqueness violations cause systematic accuracy errors in any aggregate analysis.
Sohovi scores your dataset against your own accuracy standards and highlights the columns and rows where values fall outside expected ranges.
Using These Relationships for Root Cause Analysis
When you find a quality problem, trace its dependencies:
- Consistency violation → check accuracy in both systems
- Uniqueness violation → check which record is accurate before merging
- Validity violation → confirm whether the underlying value is also inaccurate
- Completeness improvement → recheck accuracy after filling in missing values
Treating dimensions in isolation leads to partial fixes. Understanding dependencies leads to complete ones.
