Non-Invasive Household Sensor Dataset — Motion, Doors/Windows & Temperature [Multi-Home]
Catalog Summary
"Real smart-home IoT dataset from non-invasive PIR motion, magnetic door/window, and temperature sensors installed in multiple households. CSV format (6.4 MB). Used for occupancy detection, activity recognition, and smart home automation research."
Catalog Notes
Overview
This dataset originates from non-invasive IoT sensors — passive infrared (PIR) motion detectors, magnetic contact sensors on doors and windows, and ambient temperature sensors — installed in real residential households. The non-invasive nature of the sensors is a key design feature, preserving occupant privacy while still enabling robust models for occupancy detection, resident behavior monitoring, and home automation control.
The training and testing splits were prepared for machine learning experiments focused on activity recognition and occupancy inference, using only low-bandwidth binary and scalar sensor events rather than cameras or microphones. This approach reflects the growing deployment model of privacy-preserving smart home IoT systems in Europe and elsewhere.
The dataset is hosted on Zenodo (published March 2025) and is directly downloadable as a single ZIP archive of 6.4 MB, making it lightweight and easy to work with for rapid prototyping and benchmarking of smart home analytics algorithms.
Column Schema
| Column | Description |
|---|---|
| timestamp | Date and time of the sensor event. |
| sensor_id | Unique identifier for the individual sensor. |
| sensor_type | Type of sensor: motion (PIR), door, window, or temperature. |
| location | Room or zone where the sensor is installed (e.g., living room, hallway). |
| value | Sensor reading: binary (0/1) for motion/door/window; numeric for temperature. |
| household_id | Anonymized identifier for the monitored household. |
| split | Train or test set designation. |
Key Statistics
- Sensor Types: PIR motion, magnetic door/window contact, ambient temperature
- Households: Multiple anonymized residential households
- File Format: CSV (within ZIP archive)
- File Size: 6.4 MB (compressed ZIP)
- Published: March 2025 on Zenodo
- Split: Pre-defined training and testing partitions
Use Cases
- Occupancy detection and resident presence inference using non-invasive sensors
- Activity of daily living (ADL) recognition with privacy-preserving IoT data
- Smart home automation trigger modeling based on door, window, and motion events
- Lightweight edge ML model development for resource-constrained IoT gateways
Source & Attribution
The dataset was published to Zenodo in March 2025. It is fully open and downloadable without registration, making it well-suited for academic and research use in smart home IoT, ambient assisted living (AAL), and intelligent building automation studies.
View Data Structure
To explore column names, data types, and sample rows, visit the official dataset page on Zenodo.
Preview on ZenodoCite This Dataset
Llumiguano, H., Santofimia, M. J., del Toro Garcia, X., Bolaños, C., Fernández-Bermejo Ruiz, J., VILLANUEVA MOLINA, F. J., & Rocha, P. (2025). Non-invasive sensor data in households (motion, doors/windows, and temperature). [Dataset]. Zenodo. https://doi.org/10.5281/ZENODO.14832143
Source metadata: Zenodo (2025) · DOI: 10.5281/ZENODO.14832143
Indexed by IoTDataset.com on Apr 17, 2026
Review the Source Record
Confirm the licence, version, access conditions, file format, and provenance at the source before use.