Dataset: AI Usage
Entity: AI Usage Measurement
Field ID: language
Type: Text
Description: The programming language the measurement relates to. Values are always lowercased, so python never appears beside Python and a report does not need to case-fold. Only present on LANGUAGE_SURFACE and LANGUAGE_MODEL records.
Note: GitHub Copilot is the only connected app that attributes usage to a language. A language report over any other app is empty by design, not broken.
Source: App
Transformation logic: The provider's value is lowercased. Nothing else about it is changed, so a language Keypup has never seen still arrives intact.
App Mapping
GitHub Copilot | the language published on Copilot's per-language reports, lowercased |
Anthropic Claude (coming soon) | N/A - Anthropic attributes no usage to a language |
Cursor (coming soon) | N/A |
OpenAI Platform (coming soon) | N/A |
OpenAI Codex (coming soon) | N/A |
Reporting Use Cases
The Language field shows which parts of a codebase AI tooling is actually being used on.
Adoption by technology: Group by language over the language breakdowns to see where AI assistance is concentrated, and where it is not being adopted. Because values are lowercased, one language is always one group.
Filter on the breakdown first: This field is only populated on
LANGUAGE_SURFACEandLANGUAGE_MODELrecords. Without that filter, most records return an empty language.No prompt counts on language records: Copilot's per-language reports publish code measures only, so
interactionsis always empty onLANGUAGE_SURFACEandLANGUAGE_MODELrecords. A language report can show suggestion and lines-of-code volume; it cannot show prompt volume.GitHub Copilot only: No other connected app attributes usage to a language, so a language report scoped to another app returns nothing.