Tag Weight Modifier¶
Every tag carries a Weight Modifier, a whole number from -10 to 10 that says how much the assets carrying the tag matter. New tags start at 1. The weight works in two places: it changes how much a table or field counts in its quality score, and it raises or lowers the importance of the checks that carry the tag.
| Weight | Use it for | Effect on check importance |
|---|---|---|
-10 to -1 |
Data that should count less, such as test, sandbox, or deprecated tables. | One level down. |
0 or 1 |
Informational labels that should not change anything. | No change. |
2 to 7 |
Data that matters more than usual, such as a core domain. | One level up. |
8 to 10 |
The data you are responsible for, or that the business depends on. | Two levels up. |
Hover the help icon next to Weight on a tag's page, or in the tag form, to see the same bands for the weight you picked.
How Tag Weights Affect Quality Scores¶
A datastore's quality score is a weighted average of its tables and files, and a table's dimension scores (completeness, conformity, and the others) are a weighted average of its fields. Tag weights decide those weights.
At the datastore level, each table or file gets a weight from the tags it carries, including the ones it inherits from the datastore:
Table weight = sum of the weight modifiers of its tags, shifted so the lowest weight in the datastore is 1
Example: a datastore has 50 tables. You tag 3 of them with Monitored (weight 10) and leave the other 47 untagged. Each tagged table gets a weight of 11, and each untagged one a weight of 1. Once their weights are recalculated, the 3 tagged tables make up about 41% of the datastore score instead of 6%. See When Weights Are Recalculated.
At the table level, fields work the same way: fields with higher-weight tags count more in the table's dimension scores.
Because weights are shifted so the lowest one is 1, a negative weight does not remove a table from the score. It makes the table count less than its untagged neighbors.
Score only what you own
To see a score for the tables you own alone, give them a tag and open the tag's page: its Quality Score card shows the score across the tables and files that carry it. You can also filter Explore by the tag. A high weight on that tag also makes those tables count more in the datastore score, once their weights are recalculated.
See Data Quality Score for how the scores themselves are built.
When Weights Are Recalculated¶
Each table and field keeps a stored weight, and only some tag changes recalculate it. Until then, quality scores keep using the previous weight.
| Change | Table weights in the datastore score | Field weights in a table's scores |
|---|---|---|
| A tag is added to or removed from a datastore | Recalculated. The datastore score is refreshed once the background update finishes. | Recalculated for the datastore's tables. Each table's score reflects it from its next scan. |
| A tag is added to or removed from a table or file | Not recalculated. | Recalculated for that table's fields. Its score reflects it from its next scan. |
| A tag is added to or removed from a field | Not recalculated. | Not recalculated. |
| Move to, Remove, or Merge from a tag's page | Recalculated for the tables the action reaches. The datastore score is refreshed right away. | Recalculated for the fields the action reaches. |
| A tag's weight is edited | Not recalculated. | Not recalculated. |
| A tag is deleted | Not recalculated. | Not recalculated. |
Check importance and the impact of open anomalies work differently: they update on their own after each of these changes.
How Tag Weights Affect Check Importance¶
A check's Importance (Low, Medium, High, or Critical) decides how much its failures count when Qualytics ranks the impact of the anomalies it finds. Tag weights move it:
- Only the tag with the highest weight raises importance:
2to7raises it one level,8or more raises it two levels. - A tag with a negative weight lowers it one level. Several negative tags still lower it only once.
- Tags weighted
0or1have no effect.
Tags a check inherits from its table or datastore count too. Tags on an anomaly itself don't change its impact: only the tags on its failed checks do. When a tag's weight changes, the importance of the tagged checks, and the impact of their open anomalies, update on their own.
See Check Importance for the full rules and examples, and Impact for how importance feeds an anomaly's impact.
Weights on Team and External Tags¶
A team tag carries a weight like any other tag. Since it only goes on its team's datastores, its weight only moves scores there.
An external tag takes its name and color from the data catalog, but its weight is set in Qualytics, so it counts in scores and importance like any other tag.
See Also¶
-
Global Tags
Tags open to the whole workspace, managed by Admins.
-
Team Tags
Tags scoped to one team, and what happens when the scope changes.
-
External Tags
Tags imported from a connected data catalog.
-
Tag Page
The summary, timeline, assets, and map of a single tag.
-
Examples
Real scenarios showing tags at work.
-
Best Practices
Guidelines for a small, consistent tag library.
-
Permissions
Who can create, apply, and manage tags.
-
Introduction
What tags are and why to use them.
-
How It Works
How tags influence Qualytics, the three types, and how a tag spreads through a datastore.