fitness-tracker-dataset

# 🏃 Synthetic Wearable & Activity Dataset — Powered by Syncora.ai **Free dataset** for health analytics, activity recognition, **synthetic data generation**, and **dataset for LLM training**. --- ## 🌟 About This Dataset This dataset contains **synthetic wearable fitness records**, modeled on signals from devices such as the Apple Watch. All entries are **fully synthetic**, generated with **Syncora.ai’s synthetic data engine**, ensuring **privacy-safe and bias-aware** data. The dataset provides rich **biometric and activity metrics** — from steps and heart rate to entropy measures and resting heart rate — enabling experimentation in **time-series ML, predictive health analytics, and AI-driven wellness research**. **Visit our website** to learn how to **generate synthetic data** for your own projects: [🌐 Syncora.ai](https://syncora.ai) --- ## 📊 Dataset Features | Feature | Description | |---------|-------------| | **age / gender** | Demographic attributes (integer encoded gender: 0 = female, 1 = male) | | **height / weight** | Biometric profile in cm/kg | | **steps / distance** | Daily step count and distance traveled | | **heart_rate / resting_heart** | Active & resting heart rate signals | | **calories** | Estimated calorie expenditure | | **entropy_heart / entropy_steps** | Information entropy of biometric signals | | **corr_heart_steps** | Correlation between steps and heart rate | | **intensity_karvonen** | Exercise intensity computed via Karvonen formula | | **device** | Wearable source (Apple Watch in this dataset) | | **activity** | Activity label (e.g., Lying, Walking, Running, etc.) | --- ## 📦 What’s Inside - **Synthetic Wearable Dataset (CSV)** – Ready for download [⬇️ Download Dataset](https://huggingface.co/datasets/syncora/fitness-tracker-dataset/blob/main/wearables_monitoring_data.csv) - **Jupyter Notebook** – End-to-end ML example [📓 Open Notebook](https://huggingface.co/datasets/syncora/fitness-tracker-dataset/blob/main/apple_watch_fitbit_data_synthetic%20(1).ipynb) ## 🔗 Resources - **Generate Your Own Synthetic Data** – Build datasets with our generator [⚡ Synthetic Data Generator](https://huggingface.co/spaces/syncora/synthetic-generation) - **Learn More** – Visit our website [🌐 Syncora.ai](https://syncora.ai) --- ## 🤖 Machine Learning & AI Use Cases - **🏃 Activity Recognition**: Train ML models to detect physical states (e.g., lying, walking, running) - **🩺 Health Monitoring**: Develop synthetic baselines for heart rate and calorie burn models - **📊 Time-Series Forecasting**: Predict steps, calories, or heart signals over time - **🧠 LLM Training**: Use structured-to-text conversion for **dataset for LLM training** - **⚡ Synthetic Data Validation**: Compare models on real vs **synthetic data generation** - **🔍 Explainability & Feature Engineering**: Apply SHAP/LIME for health signal interpretation --- ## 🚨 Why Synthetic? - **Privacy-Safe**: No real user data, safe for open research - **Bias Reduction**: Balanced synthetic sampling enables fairer AI models - **Rapid Prototyping**: Quickly test models without waiting for costly real-world data - **Flexible Expansion**: Generate synthetic data at scale with [Syncora.ai](https://syncora.ai) --- ## 📜 License Released under **MIT License**. This is a **100% synthetic, free dataset** — ideal for **AI research, synthetic data generation, and dataset for LLM training**.