Hi Product Hunt 👋

I built PulseClean to make wearable health data easier to clean and use for analysis.

Right now, PulseClean supports Apple HealthKit XML and generic wearable CSV files. It converts messy exports into a standardized CSV with quality flags, conservative missing-value handling, outlier handling, and documentation for the preprocessing policy.

The default pipeline is designed to avoid future-value imputation, so it is safer for ML-oriented workflows than simple bidirectional interpolation.

PulseClean is still early, and I would really appreciate feedback from researchers, data scientists, builders, and anyone working with wearable data.

If a preprocessing rule does not fit your workflow, or if you need support for another wearable format, please let me know. I’ll prioritize updates based on real user needs.

Thanks for checking it out.