Tags are a way for attaching arbitrary data to your rows so they can be sliced and diced.

This section provides a more granular look at tags, the bits of metadata we attached to our predictions in the last section to understand how they impact users. The mental model for tags in a Gantry row is just additional columns:


More Detail

A Tag is an arbitrary key/value pair that is attached to a record at ingestion time to facilitate identifying that record via filters. Examples of tags might include:

  • Which environment was a record was generated in (prod, sandbox, dev)

  • Which model version generated the prediction that a record represents (for example a code or data hash, or combination)

  • Other arbitrary data, such as the end consumer of the prediction for cases where a single model service serves multiple user interfaces

Tags are created by populating the tags keyword argument in the Python SDK, or via tags argument in our REST API. For example the following dictionary (or similar JSON), would tag the associated record served in production to an iOS application:

    "env": "prod",
    "version": "c09f689b93facc1f0e165b0d98f9a9cd52de0668",
    "consumer": "mobile-ios"

n summary tags are flexible way to tag your data with descriptive information about where predictions come from and the context predictions are served in. The data type of all tags is string and cannot be changed. Filtering on substrings works and tags can be searched for in the UI: