You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
{{ message }}
This repository was archived by the owner on Oct 3, 2026. It is now read-only.
Repository navigation
This repository was archived by the owner on Oct 3, 2026. It is now read-only.
Are you able to provide a diagram or an overview of the data schema for Soundscape (including that which was handled on the phone and that which was handled by the services…and the underlying data structure in support of this)
Peak active users at any given time
Any other telemetry you can share which would be useful
I can provide information that might inform anybody bringing up the service now. The current production service after the announcement still has a few hundred unique users a day with about 5 times that for a 7 day week. The app has substantial caching functionality so that limits some services impact at least in the production configuration.
Unfortunately I can't provide telemetry. The user agreed to provide that telemetry to Microsoft and not a 3rd party. There have been some past research projects that used Soundscape -- perhaps they've published something that would help.
I can't provide a direct schema but i can provide a description of the division of storage/state in the open source version. The iOS client stores markers , routes, and all user preferences. It also stores cached OSM data as retrieved from the service. This is all stored in a realm database.
In contrast, the service retrieves OSM data from one of the OSM mirrors, processes it using IMPOSM3 -- storing it in a Postgres database. Upon tile request to the service (in sources as 'gentile.py', this service uses a PostGIS sql query to retrieve the data that applies to that tile and transforms it to GeoJSON.
There are no acccounts, no PII storage, no concept of individual users as seen by the core service, and no sharing via a centralized service.
There was an additional web service that was used for some web authoring scenarios but consuming one of those routes, just inserted it into the local storage.