ADR-0038 UC3 Recommender service no longer needs analytics user events

Publication Date2022-06-27
Last Update2022-06-27
StatusProposed

Context

In ADR 0024 a decision was made for user events to be saved in Firebase Analytics and then be subsequently read by the Recommender service.

Since then it has been made clear that all user events that are saved in Firebase Analytics would not be useful to the recommender service.

Thus we don’t need to share any user events through Firebase Analytics and subsequently Big Query.

The only data that the Recommender service has use of are users favorites, that are sent to the service through an API call from inside the Magic Mirror app.

Decision

We won’t save custom user events to feed to the Recommender service, because all events that can be saved are of no use to the system.

Consequences

With this change, the only customized user data that the recommender has access to, are each user’s favorites that are fed to the recommender service through an API Request from inside the app.