Dataset: AI Usage
Entity: AI Usage Measurement
Field ID: actor_username
Type: Text
Description: The name of the person the measurement is attributed to. Note that names are app specific: GitHub reports a login while other providers report the person's display name, or their email address when they have none. API keys and service accounts are reported under their name or ID.
Note: Empty on REPOSITORY, ORGANIZATION and ENTERPRISE records, which have no single actor.
Source: App
Transformation logic: N/A
App Mapping
GitHub Copilot |
|
Claude Platform | the member's name, else |
Cursor (coming soon) |
|
OpenAI Platform | the member's name, else their email address, for |
OpenAI Codex Enterprise | the user's name, else their email address, as published by the analytics API |
Reporting Use Cases
The Actor field is the basis of all per-developer AI reporting: who is using AI tools, how much, and on what.
Adoption by person: Use it as a dimension with
SUM(interactions)orSUM(output_tokens)to see how AI usage is distributed across the team.Active developer counts:
COUNT_DISTINCT(actor_username)overaggregation_level = USERrecords gives the number of people who used an AI tool in a period, on one consistent definition across providers.Personal dashboards: Use the is me operator to build a widget that shows only the viewer's own AI usage.
Excluding service accounts: Filter on
actor_type = HUMANrather than pattern matching on the username.Not comparable across providers: GitHub gives a login and the other providers give the person's name or email address, so the same person appears as two actors when two integrations are connected. Scope cross-provider per-person reports to one provider, or expect the distinct count to be too high.