Cohesion
Artsy's analytics schema & event helpers
Contributing
Set up:
yarn install
Run tests:
yarn test
Generate docs:
yarn docs
Schema
The /Schema
directory represents the Artsy's analytics schema, and describes expectations for data consumed by Redshift and Segment.
This schema is maintained by Artsy's data team, engineers should not expect to change these files.
Valid analytics events are described in /Schema/Event.ts
.
Typings for all allowed values, such as ContextModule
, are exported for use by engineers in consuming projects.
Add a new event to the schema (For data analysts)
- In
Schema/ActionType.ts
, add the name of the new event. This name defines the corresponding downstream table's name in Redshift, and should use the lowerCamelCase
naming convention.
export enum ActionType {
...
myNewEvent = "myNewEvent",
...
}
- In
Event.ts
, create an interface describing the shape of the new event, as it is recieved in Segment/Redshift.
The action
key is required and should match the ActionType
created in step 1.
export interface MyNewEvent {
action: ActionType.myNewEvent
context_module: ContextModule
optional_property?: string
required_property: number
}
-
If your event uses values not yet in the schema, such as a new ContextModule
, add new values to existing enums in the Schema directory.
-
Add descriptive comments with examples to explain the use of your new event. Our documentation is generated automatically from in-code comments, find more information on syntax in the typedoc
docs.
-
PR your changes. Once merged, the schema will be updated and your new event and values will be available to consumers of this package.
-
Data analysts should request an engineer to construct a new event helper for the /Events
directory (see below).
Events
The /Events
directory contains javascript helpers that return schema-compliant analytics events, and provide some useful default values. Each helper corresponds to one event from /Schema/Events.ts
.
Engineers should use these helpers whenever sending analtics data to Segment, for example, when creating and analytics event with react-tracking
.