Project
Athlete performance analytics — wearable telemetry streamed live into a bold public product.
What it is
AllSports Monitor is a sports performance platform: GPS and heart-rate wearables feed live sessions that athletes and coaches watch as they happen, then dig into afterwards. It is a full consumer product — public site, subscription tiers, a shop, and team views — not an internal dashboard.
Architecture
Telemetry starts at an edge layer of GPS pods and heart-rate monitors, syncs through an offline-first mobile app over Bluetooth, and enters the backend through an API gateway. Behind it sit focused microservices — users, devices, analytics, teams, notifications — with time-series data in TimescaleDB, events on Kafka, Redis for hot state, and object storage for bulk session data.
Realtime under a public UI
Live sessions stream to the browser over WebSockets, so a coach's screen follows the pitch rather than a polling interval. The constraint that shaped the whole platform: all of that realtime machinery has to hold up behind a public consumer UI with sign-ups, pricing, and a shop — not behind a controlled enterprise login.
Notable engineering
The time-series pipeline is the heart of it: high-frequency positional and biometric samples are ingested and kept for post-session analysis while live views stream as the session happens. Metrics are user-definable rather than hard-coded, and the stack is built for data portability — a team's sessions can leave the platform in open formats.
Have realtime data to tame?
Streams, sensors, and dashboards that have to stay honest under load — tell me what you're measuring.