View asset quality scores - Precisely Data Integrity Suite

View quality scores for your datasources, datasets, and fields. Understand how scores are calculated and use them to identify data quality issues. Quality scores measure how well your data meets defined quality standards. Scores aggregate factors such as freshness, completeness, validity, and accuracy. Higher scores indicate assets you can trust for critical operations. Scores flow hierarchically from individual rule evaluations up through fields, datasets, and datasources. This structure helps you identify issues and track improvements.

1. Go to Catalog and select the Datasources, Datasets, or Fields tab.
2. Quality scores display in both card and list views.
   • Card view: scores appear next to the asset name
   • List view: a Quality Score column shows scores for each row
3. Select an asset and go to the Scoring tab to view dimensions and associated rules.

   The Scoring page displays the overall score and quality score dimensions for the selected asset.

4. Each level aggregates the scores below it:

   Score Level	Formula	Example	Key point
   Rule Score	(Records passing / Total records) × 100	950 passing out of 1000 = 95%	Ranges from 0% to 100%
   Field Score	Average of all rule scores applied to the field, along with associated dimensions.	Rules: 98%, 96%, 94% = 96%	All rules weighted equally
   Dataset Score	Average of all field scores in the dataset, along with associated dimensions	Fields: 99%, 96%, 92%, 88% = 93.75%	Only fields with scores included
   Datasource Score	Average of all dataset scores within that datasource, along with associated dimensions	Datasets: 93.75%, 91.5%, 95.2% = 93.48%	Represents overall data health

   Tip:

   • When using functions such as Not Equal, Less Than, or Less Than or Equal, null records are included in the calculation. To exclude null values, add a secondary condition using the 'is not null' operator.
   • Fields with lower scores proportionally impact the overall dataset quality assessment.
5. Calculate scores when a field has multiple rules across different dimensions.

   Each rule generates its own score. Rules are grouped by dimension, and dimension scores are calculated as the average of all rules within that dimension. The field score is the average of all dimension scores. For example, a customer email field has rules across two dimensions:

   • Completeness: Email not null = 98%
   • Validity: Valid email format = 96%, Email domain exists = 94%
   • Validity score = (96 + 94) / 2 = 95%
   • Field score = (98 + 95) / 2 = 96.5%
6. Calculate dataset scores when only some fields have rules assigned.

   Only fields with assigned rules contribute to the dataset score. Fields without rules are excluded. For example, a customer dataset has 10 fields, but only 6 have quality rules with scores of 99%, 96%, 92%, 88%, 94%, and 91%:

   Dataset Score = (99 + 96 + 92 + 88 + 94 + 91) / 6 = 93.33%

   Tip: Assign rules to all critical fields to ensure comprehensive quality assessment.
7. Know the quality score dimensions that determine how aggregate scores are calculated.

   Dimension	Definition
   Completeness	Assesses whether required data is present and not missing. For example, rules checking for null or blank values. If 98 out of 100 records are not null, the completeness score is 98%.
   Validity	Checks whether data conforms to predefined formats, standards, or constraints. If the semantic type confidence is 98%, the validity score is 98%.
   Field Score	The aggregate score of all quality dimensions for a field.

8. Track when quality scores update based on rule execution. Quality scores update when rules run. The timing depends on how rules are configured:

   Update Type	Behavior
   Manual runs	Scores update immediately when you run a rule from the Rules page or rule details page.
   Scheduled runs	Scores update according to the rule's configured schedule. By default, rules follow the discovery schedule.
   Cascading updates	When a rule score updates, all dependent scores (field, dimension, dataset, and datasource) automatically recalculate.

   The timestamp on the Scoring page shows when the most recent rule evaluation occurred.

9. Interpret quality score ranges to assess data quality. Quality scores measure data quality on a scale of 0% to 100%:

   Score Range	Assessment
   90–100%	Excellent quality. Data is suitable for critical business operations and analytics.
   80–89%	Good quality. Data is generally suitable for most operations but may require attention in specific areas.
   70–79%	Fair quality. Data has notable quality issues that should be addressed before use in critical operations.
   Below 70%	Poor quality. Data requires significant remediation before use in production systems.

   Tip: Establish your own quality thresholds based on business requirements and risk tolerance.