---
license: mit
task_categories:
- time-series-forecasting
- sequence-modeling
- reinforcement-learning
language:
- en
tags:
- smart-home
- iot
- spa
- device-prediction
- time-series
- user-behavior
- kohler-anthem
- smart-spa
size_categories:
- 1K<n<10K
pretty_name: "Smart Spa Usage Dataset"
configs:
- config_name: sessions
  data_files: "sessions.json"
- config_name: events  
  data_files: "events.json"
- config_name: sequences
  data_files: "sequences.json"
---

# Smart Spa Usage Dataset

## Dataset Description

This dataset contains real usage patterns from a smart spa system based on Kohler Anthem+ technology. It includes comprehensive user interaction data, device usage sequences, environmental parameters, and temporal patterns collected for machine learning research in smart home automation and predictive device control.

The dataset captures natural user behavior patterns in a smart spa environment with 20 different devices including shower systems, steam generators, lighting controls, audio systems, and various wellness programs.

## Dataset Structure

### Data Instances

The dataset consists of three main components:

#### 1. Sessions Data (`sessions.json`)
- **Total Records**: 1,000 unique spa sessions
- **Fields**: Session metadata, user preferences, timing information
- **Format**: JSON Lines (.jsonl)

#### 2. Events Data (`events.json`) 
- **Total Records**: 9,353 individual device interaction events
- **Fields**: Device activations, environmental parameters, temporal data
- **Format**: JSON Lines (.jsonl)

#### 3. Sequences Data (`sequences.json`)
- **Total Records**: 4,417 processed sequences for ML training
- **Fields**: Sequential patterns, target predictions, feature vectors
- **Format**: JSON Lines (.jsonl)

### Data Fields

#### Sessions
- `session_id`: Unique session identifier
- `user_id`: Anonymous user identifier  
- `start_time`: Session start timestamp
- `end_time`: Session end timestamp
- `event_count`: Number of events in session
- `total_duration`: Session duration in seconds
- `session_type`: Type of spa session
- `preferences`: User preference settings

#### Structure
smart-spa-usage-dataset/
├── README.md (the content above)
├── sessions.json (1,000 sessions)
├── events.json (9,353 events)
└── sequences.json (4,417 ML sequences)

#### Events
- `session_id`: Link to parent session
- `event_order`: Sequential order within session
- `title`: Device/program name (20 unique devices)
- `status`: Device status (ON/OFF/ACTIVE)
- `start_time`: Event timestamp
- `zone1_temp`: Temperature zone 1 (°F)
- `zone1_flow`: Water flow zone 1 (%)
- `zone2_temp`: Temperature zone 2 (°F) 
- `zone2_flow`: Water flow zone 2 (%)
- `steam_temp`: Steam temperature (°F)
- `steam_time`: Steam duration (seconds)
- `brightness`: Lighting level (0-100)
- `volume`: Audio volume (0-100)

#### Sequences
- `sequence`: Array of 5 consecutive events (12 features each)
- `target`: Next device prediction target
- `features`: Processed feature vectors for ML

### Device Categories
The dataset includes 20 smart spa devices:
- **Core Systems**: SHOWER, STEAM, LIGHT, AMPLIFIER
- **Wellness Programs**: Relax, Restore, Focus, Breathe, Detox
- **Temperature Therapy**: Advanced Ice Shower, Beginner Ice Shower, Warm Up, Cool Down
- **Steam Programs**: Steam Coach (Beginner/Medium/Advanced)
- **Experience Programs**: FAVORITE, Headaid, Repair, Flex

## Dataset Creation

### Data Collection
- **Source**: Real smart spa usage logs from Kohler Anthem+ system
- **Duration**: Extended monitoring period capturing diverse usage patterns
- **Privacy**: All user identifiers anonymized, no personally identifiable information
- **Preprocessing**: Cleaned, validated, and structured for ML applications

### Data Processing Pipeline
1. Raw spa system logs collected
2. Data cleaning and validation
3. Time-based feature engineering
4. Sequence generation for prediction tasks
5. Train/validation/test splits created
6. Export to standardized JSON format

## Use Cases

### Primary Applications
- **Device Usage Prediction**: Predict next device user will activate
- **Smart Home Automation**: Automate spa systems based on behavior patterns
- **Reinforcement Learning**: Train RL agents for optimal spa control
- **Time Series Forecasting**: Predict usage patterns and preferences
- **User Behavior Analysis**: Understand spa usage habits and preferences

### Machine Learning Tasks
- Sequential pattern recognition
- Multi-class classification (20 device classes)
- Time series prediction
- Reinforcement learning environment
- Anomaly detection in usage patterns

## Model Performance Baselines
Baseline results from testing various ML approaches:
- **LSTM Model**: 42-45% device prediction accuracy
- **DQN Reinforcement Learning**: 18-60% accuracy (varies by implementation)
- **Random Baseline**: ~5% accuracy (1/20 devices)

## Considerations for Using the Data

### Limitations
- Data collected from single spa system type (Kohler Anthem+)
- Limited to 20 specific device types
- Temporal patterns may be specific to data collection period
- User base may not represent all demographics

### Ethical Considerations
- All personal identifiers removed
- Usage patterns anonymized
- No health or personal information included
- Suitable for research and commercial applications

### Recommended Use
- Smart home automation research
- IoT device prediction studies
- Reinforcement learning experimentation
- Time series analysis education
- Spa/wellness technology development


## Citation

If you use this dataset in your research, please cite:

@dataset{smart_shower_usage_2025,
title={Smart Spa Usage Dataset: Real-world IoT Device Interaction Patterns},
author={[Zachzzz33]},
year={2025},
publisher={Hugging Face},
url={https://huggingface.co/datasets/Zachzzz33/smart-shower-usage-dataset}
}

## Additional Information

## Contact & Support

For questions, issues, or contributions to this dataset:
- Open an issue in this repository
- Contact: [sachinsachu9500@gmail.com]
- Research collaboration inquiries welcome
  
### License
This dataset is released under the MIT License, allowing for both research and commercial use.

### Acknowledgments
Data collected from Kohler Anthem+ smart spa system. Thanks to the spa technology community for supporting open research datasets.

**Last Updated**: September 2025  
**Dataset Version**: 1.0  
**Maintainer**: [Zachzzz33/Sachin]