Below is a collection of resources that provide accelerometer data. This is a compilation of datasets from our group and others. We have also begun to identify datasets that provide data from multiple sensor types and would welcome feedback about this curated list of movement-related datasets.
To find other movement-related data resources, such as from motion capture systems or force sensors, visit sites such as Simtk.org and PhysioNet. If you know of other resources that should be on this list, please let us know.
Data Source:
Understanding Walking in Parkinson’s Disease with Wearable Sensors
Entity Supplying Data:
Bronte-Stewart lab on Human Motor Control and Neuromodulation, Stanford University
Description:
This is data collected from 20 subjects with Parkinson’s disease who experience severe walking symptoms like freezing of gait (FOG), also identified as (freezers); as well as data from people with Parkinson’s disease who do not experience FOG (non-freezers). The dataset also includes a small number of age-matched healthy control subjects. The data provides gait parameters calculated from wearable measurement unit sensors while subjects walk in a turning and barrier course, as well as in a straight path along a hallway. Demographics and their Freezing of Gait Questionnaire scores are also included.
License:
See website for citation and license agreement
Data Source:
Daphnet Freezing of Gait Data Set
Entity Supplying Data:
Laboratory for Gait and Neurodynamicsm, Tel Aviv Sourasky Medical Center, Israel and the Wearable Computing Laboratory, ETH Zurich, Switzerland
Description:
This is a dataset devised to benchmark automatic methods to recognize gait freeze in individuals with Parkinson’s disease from wearable acceleration sensors placed on the legs and hip. The dataset was recorded in the lab with an emphasis on generating many freeze events. Subjects performed three kinds of tasks: straight line walking, walking with numerous turns, and activities of daily living (ADL), such as going into different rooms while fetching coffee, opening doors, etc.
License:
See website for citation
Data Source:
The Osteoarthritis Initiative (OAI) dataset
Entity Supplying Data:
OAI Consortium
Description:
This is a multi-center, longitudinal, prospective observational study of knee OA that is publicly available (https://oai.epi-ucsf.org). The data include measurements from biospecimens (serum, plasma, and urine), X-ray and MR imaging, clinical exam notes, and questionnaire responses. At the 48-month mark, 7-day accelerometry data were also collected for 2712 of the original 4796 participants.
License:
Data use agreement
Data Source:
The National Health and Nutrition Examination Survey (NHANES)
Entity Supplying Data:
National Center for Health Statistics
Description:
This is a large continuous study designed to assess the health of the U.S. population, providing samples that are representative of the U.S. population. Physical activity monitor data collected in 2003-2004 and 2005-2006 for participants 6 years and over is available. Physical activity monitor data collected in 2011-2012 and 2013-2014 for participants 3 years and over is also available.
License:
Data use agreement
Data Source:
Human Activity Recognition Using Smartphones Data Set
Entity Supplying Data:
Smartlab – Non-Linear Complex Systems Laboratory DITEN – Università degli Studi di Genova and CETpD – Technical Research Centre for Dependency Care and Autonomous Living Universitat Politècnica de Catalunya (BarcelonaTech)
Hosted on the UCI Machine Learning Repository
Description:
This data set contains Activities of Daily Living (ADL) recordings for six activities that include walking, walking upstairs, and standing from 30 participants ranging from 19-48 years of age. The data data was recorded through a waist-mounted smartphone with embedded inertial sensors.
License:
See website for citation
Data Source:
ADL Recognition with Wrist-worn Accelerometer Data Set
Entity Supplying Data:
Laboratorium – Laboratory for Ambient Intelligence and Mobile Robotics
DIBRIS, University of Genova
UCI Machine Learning Repository
Description:
This data set contains 14 Activities of Daily Living (ADL) recordings of simple activities such as brushing teeth, climbing stairs, and drinking water from 16 volunteer participants. The data was recorded through a right-wrist worn tri-axial accelerometer on each of the participants.
License:
See website for citation
Data Source:
Activity Recognition from Single Chest-Mounted Accelerometer Data Set
Entity Supplying Data:
Computer Vision Center, Bellaterra, Stain
University of Barcelona, Spain
UCI Machine Learning Repository
Description:
This data set contains uncalibrated accelerometer data from 15 participants. The data was recorded using an accelerometer mounted on the chest and recorded seven activities that included walking and standing.
License:
See website for citation
Data Source:
International Children’s Accelerometry Database (ICAD)
Entity Supplying Data:
University of Cambridge MRC Epidemiology Unit
Description:
This is data from a consortium of 20 partners that pooled and reduced raw accelerometer data using standardized methods to process the data and create variables based on over 37,000 participants ages 3 to 18.
License:
Data use agreement. Must apply to get access to data
Data Source:
UK Biobank
Entity Supplying Data:
UK Biobank
Description:
This is a study that collected health information from 500,000 people between the ages of 40-69 years old. The data was collected from 2006-2010. A week’s worth of data recorded from a 24-hour activity monitor for 100,000 participants is also available. Repeat measurements are available from 20,000 of these.
License:
Data use agreement. Must apply to get access to data
Data Source:
Long Term Movement Monitoring Database
Entity Supplying Data:
PhysioBank
Description:
This data set contains accelerometer recordings for three days from 71 older adults ages 65-87 years old.
License:
Unknown
Data Source:
Avon Longitudinal Study of Parents and Children
Entity Supplying Data:
University of Bristol
Description:
This is a longitudinal study with over 14,000 pregnant participants. These participants have been followed for 19-22 years. The dataset includes 1 to 7 days of physical activity data recorded with a uniaxial accelerometer for over 1,300 young people born from pregnant participants.
License:
Data use agreement. Must apply to get access to data
Data Source:
Multicenter Osteoarthritis Study (MOST)
Entity Supplying Data:
Multicenter Osteoarthritis Study Public Data Sharing
Description:
This is a longitudinal study with over 3,000 participants ages 50 to 79 at baseline. Participants had followup contacts at 15, 30, 60, 72, and 84 months after initial baseline assessments, which began in 2003. The dataset includes physical activity data for a subset of participants that was assessed through accelerometer collected at 60 and 84 month followups.
License:
Data use agreement. Must apply to get access to data
Data Source:
MHEALTH Dataset
Entity Supplying Data:
Department of Computer Architecture and Computer Technology, University of Granada
UCI Machine Learning Repository
Description:
This data set contains body motion and vital signs data from ten volunteers while performing 12 different physical activities. Measurements included acceleration, rate of turn, and magnetic field orientation. The data was recorded with Shimmer2 sensors placed on the participant’s chest, right wrist, and left ankle.
License:
See website for citation