Data Quality Score

The Data Quality Score is a key indicator that helps data stewards, engineers, and analysts prioritize remediation efforts and build trust in governed data assets. It reflects the overall quality of monitored data across your catalog. It provides a quick overview of how many quality checks are passing successfully and is updated daily or weekly as scheduled to reflect the most up-to-date quality status.

A higher score indicates better data quality and fewer issues detected, while a lower score highlights potential data reliability concerns that may require attention.

The Data Quality Score is designed to help users quickly assess whether data meets expectations across multiple quality dimensions like completeness, uniqueness, validity, and freshness.

The quality score appears on the following pages:

  • Catalog pages

  • Data Quality dashboards

  • Alation Chrome Extension

The health score calculation is an indicator of the number of active checks that have passed over the total number of enabled active checks indicating the following:

  • Executed checks include only those with results of pass or fail.

  • All checks contribute equally.

Score Calculation

Health Score = (Number of Passed Active Checks / Total Number of Enabled Active Checks) x 100

Score Interpretation Guidelines

  • 90-100%: Excellent Excellent data quality with minimal issues (Green)

  • 70-89%: Good Good quality with some attention required (Yellow)

  • Below 70%: Poor Poor quality requiring immediate attention (Red)

Example

Consider tables customer_profiles and orders with 5 data quality checks:

Table

Check

Result

customer_profiles

missing_percent(email) < 2%

Fail

customer_profiles

duplicate_count(customer_id) = 0

Pass

customer_profiles

freshness(updated_at) < 1d

Pass

orders

row_count > 1000

Pass

orders

invalid_count(phone) = 0

Pass

Data Quality Score = (4 / 5) x 100 = 80%

To determine action, consider the following:

  • Warning or Medium Score (75-89%): Some checks have failed

  • Failing or Low Score (<75%): Significant issues are detected in the monitored data

  • 0% Score: All checks have failed and requires an immediate review