| language: en | |
| license: openrail | |
| tags: | |
| - medical | |
| - time-series | |
| - wearables | |
| - menstrual-health | |
| pretty_name: CycleBench Wearable Symptom Log Dataset | |
| size_categories: | |
| - 1K<n<10K | |
| # CycleBench Wearable Symptom Log Dataset | |
| This dataset contains anonymized, multi-modal symptom and wearable logs collected for menstrual health research. | |
| It is divided into client-specific files to allow conflict-free updates from federated learning instances. | |
| ## Dataset Structure | |
| Each file in the `data/` directory belongs to a unique client (`data/client_[UUID].csv`) containing daily tracking rows: | |
| * **Hormone levels & symptoms**: Headaches, cramps, bloating, etc. (mapped to $0$-$5$ ordinal levels). | |
| * **Wearable logs**: Resting HR, sleep minutes, sleep efficiency, steps, and wrist temperature delta. | |
| * **Target phase**: Inferred cycle phase (`Menstrual`, `Follicular`, `Fertility`, `Luteal`). | |