SPORTS DATA / ANALYTICS
ATHRA
Performance starts with understanding the signal.
01 / THE PROBLEM
From training data to a clearer picture.
Training produces a lot of data. Interpreting activity, sleep and recovery together should not require jumping between disconnected tools.
02 / THE APPROACH
How the system fits together.
ATHRA imports Garmin FIT files, normalizes activity records and calculates versioned, deterministic recovery and training-load indicators. The interface brings those calculations into focused views for daily readiness, activities, sleep and trends.
Garmin FIT files → validated upload
FIT SDK → normalized activity records
Repository interface → in-memory implementation
Recovery engine → Express API
React + Recharts → athlete dashboard
03 / ENGINEERING DECISIONS
Keep the calculation inspectable
The recovery engine computes HRV baselines, resting heart-rate deviation and exponentially weighted training load in ordinary TypeScript. The calculation is separate from the optional AI coaching integration.
Preserve the provenance
The importer and data explorer retain a route back to imported activity records. A repository interface separates ingestion, calculations and presentation.
CURRENT SCOPE
The current checkout uses an in-memory repository. Durable persistence and production authentication are further work. Recovery indicators are product calculations, not clinically validated conclusions.