Skip to content

Datastore-Level Quality Scoring

DO Calculating quality scores

The datastore quality score sits at the top of the score hierarchy. It combines the container scores of a source datastore into a single number, so you can compare data health across datastores at a glance.

How the Datastore Score Is Calculated

The datastore quality score is a weighted average of all scored containers:

Datastore Score = SUM(container_score × container_weight) / SUM(container_weight)

Only containers with a score participate. Containers that have not been both profiled and scanned are excluded from both the numerator and the denominator. See Which Containers Get a Score? for every container state.

How Tags Affect the Datastore Score

Each container's weight in the aggregate formula is derived from its tags:

container weight = sum(tag weight modifiers on this container), normalized so the minimum is at least 1
  • A container with no tags has weight = 1 (default)
  • Tags with positive weight modifiers increase a container's influence on the datastore score
  • Tags with negative weight modifiers decrease a container's influence

Example: If you tag your 2 most important tables with a Critical (weight: +10) tag, each gets weight ~11 while untagged tables remain at weight 1. Your important tables now dominate the aggregate.

For the modifier range and when container weights are recalculated, see Tag Weight Modifiers.

Historical Daily Scores

Daily datastore scores use the last 10 unique dates with quality score activity. For each date, each container's latest score up to that date is used (not just scores from that exact day), providing a continuous trend even when not all containers are scored daily.

See Also

  • Container-Level Scoring

    Learn which containers get a score, how the 70-point baseline model combines the 8 dimensions, and how timeliness and volumetrics are measured.

    Container-Level Scoring

  • Field-Level Scoring

    See how each field is scored on completeness, coverage, conformity, consistency, precision, and accuracy, with the formula behind each dimension.

    Field-Level Scoring

  • Score Settings

    Learn how datastore settings flow to containers, and how the decay period, dimension weights, and tag weight modifiers change the score.

    Score Settings

  • Interpreting Scores

    See where scores appear, how they support decisions, and how to read them as a measure of fitness for purpose.

    Interpreting Scores

  • AI Explanations

    Learn how the AgentQ-powered summary, per-dimension explanations, and Recommendations section render on the Quality Score panel.

    AI Explanations

  • Examples

    Walk through real scenarios: a container held down by two dimensions, a datastore score lower than expected, weighting critical tables, and a settings override.

    Examples

  • Best Practices

    Keep scores meaningful: what to profile and scan, how many checks to add, when to change settings, and how to read the trend.

    Best Practices

  • Troubleshooting

    Resolve missing scores, unexpectedly low values, settings and tag changes that do not apply, and missing AI explanations.

    Troubleshooting

  • Permissions

    See which roles can view quality scores and edit scoring settings.

    Permissions

  • How It Works

    Learn what the quality score is, the key concepts behind it, the levels it is calculated at, the 8 dimensions it combines, and what triggers a recalculation.

    How It Works

  • Rule Types by Dimension

    Look up which rule types feed the Conformity, Precision, Accuracy, Volumetrics, and Timeliness dimensions, with a link to each one.

    Rule Types by Dimension