An Open Source Pipeline for 24/7 HR Data

• 08 June 2025
• fitness data,
• open source,
• python,
• docker,
• sqlite,
• api

Smart watches are ugly. Whoop wants to charge me every month. Aliexpress HR rings/ Whoop clones have questionable accuracy.

The Polar Verity Sense is worn on the bicep out of view, has an SDK, and is supposed to be more accurate.

However, there is no easy way easy way to get data out of the official apps and make correlations.

What I've built is therefore a pipeline to get this working.

The Pipeline, Step by Step

1. Getting the Data Off the Device

I collect my data through an app called Polar Recorder which is also open source.

I then use Syncthing to automatically sync the data folder from my Android to my Mac. No cables, no cloud, no hassle.

2. Watching for New Files

A Python script (the “file monitor”) that lives in Docker container watches the Syncthing directory for new or changed files, and, if unmodified for awhile, it processes that data.

3. Processing

The pipeline reduces the data (for storage purposes), and saves it into a sqlite db.

3. Display

The data is hosted by Flask, living in the same Docker container, and I then use Grafana to visualize the data.