Dark Mode
Mobile Health Body Motion Dataset
Data Science and Analytics
Tags and Keywords
Trusted By
"No reviews yet"
Free
About
This dataset is centred on Mobile Health Human Behaviour Analysis and Human Activity Recognition. It features recordings from ten volunteers of varied profiles performing twelve distinct physical activities. Sensors positioned on the chest, right wrist, and left ankle capture body motion, specifically acceleration, rate of turn, and magnetic field orientation. The chest sensor also provides 2-lead ECG measurements, which can be valuable for basic heart monitoring, checking for various arrhythmias, or observing the effects of exercise on the ECG. The data was gathered in an out-of-lab setting, allowing for the natural execution of activities. This particular dataset focuses on acceleration and gyroscope data, converted into a CSV file, making it suitable for activity classification tasks.
Columns
- alx: Acceleration from the left-ankle sensor (X axis).
- aly: Acceleration from the left-ankle sensor (Y axis).
- alz: Acceleration from the left-ankle sensor (Z axis).
- glx: Gyroscope from the left-ankle sensor (X axis).
- gly: Gyroscope from the left-ankle sensor (Y axis).
- glz: Gyroscope from the left-ankle sensor (Z axis).
- arx: Acceleration from the right-lower-arm sensor (X axis).
- ary: Acceleration from the right-lower-arm sensor (Y axis).
- arz: Acceleration from the right-lower-arm sensor (Z axis).
- grx: Gyroscope from the right-lower-arm sensor (X axis).
- gry: Gyroscope from the right-lower-arm sensor (Y axis).
- grz: Gyroscope from the right-lower-arm sensor (Z axis).
- Activity: The corresponding physical activity performed.
- subject: The volunteer identification number.
Distribution
The dataset is provided as a CSV file named
mhealth_raw_data.csv. Its size is 160.51 MB. The data includes 1.22 million records (rows).Usage
This dataset is well-suited for a variety of applications and use cases, including:
- Human Activity Recognition (HAR) model development.
- Classification tasks in machine learning.
- Building deep learning models, particularly those utilising LSTM architectures, for activity analysis.
- Research and development in health and fitness monitoring.
- Studying the effects of exercise on ECG and for basic heart monitoring purposes.
Coverage
The dataset includes data from ten diverse volunteers. It covers twelve specific physical activities:
- Standing still
- Sitting and relaxing
- Lying down
- Walking
- Climbing stairs
- Waist bends forward
- Frontal elevation of arms
- Knees bending (crouching)
- Cycling
- Jogging
- Running
- Jump front & back
Data was collected using sensors placed on the chest, right wrist, and left ankle. Each session was recorded with a video camera. The dataset is designed to generalise to common daily living activities, taking into account varied body parts, activity intensity, and execution speed.
License
CC0: Public Domain
Who Can Use It
- Data scientists and machine learning engineers for developing and testing human activity recognition algorithms.
- Researchers in mobile health, biomechanics, and human behaviour analysis.
- Students undertaking projects related to wearable sensors, signal processing, and health informatics.
- Developers creating applications for fitness tracking, elder care, or rehabilitation.
Dataset Name Suggestions
- MHEALTH Human Activity Dataset
- Wearable Sensor Activity Recognition Data
- Mobile Health Body Motion Dataset
- Human Activity & Vital Signs Data
Attributes
Original Data Source: Mobile Health Body Motion Dataset
QUALITY