This repository contains sample code, sample data, and explanatory information for working with the Aurora-BP dataset released alongside the publication of the Aurora-BP study.

The dataset includes de-identified participant information, raw sensor data aligned with each measurement, and a wide variety of features derived from sensor data. The publishing of this dataset as well as the characterization of multiple feature groups across a broad population and multiple settings are intended to aid future cardiovascular research.

Note that the data contained in this repository represent a very small sample of the full dataset, meant only to illustrate the structure of the files and allow testing with the sample code. For access to the full dataset, see the data access section below.

If you use this repository, part or all of the full dataset, and/or our paper as part of your research, please refer to the dataset as the Aurora-BP dataset and cite the publication as follows:

R.J. Mieloszyk et al. "A Comparison of Wearable Tonometry, Photoplethysmography, and Electrocardiography for Cuffless Measurement of Blood Pressure in an Ambulatory Setting," IEEE Journal of Biomedical and Health Informatics, vol. 26, no. 7, pp. 2864–2875 (2022).

• docs

  Data file descriptions, a detailed overview of the Aurora-BP Study protocol, and supplemental results not included in the Aurora-BP Study publication

• notebooks

  Sample Jupyter notebooks and environment files for basic analyses using Aurora-BP Study data

• sample

  Example data files, to run sample Jupyter notebooks and provide researchers a direct look at the data format before applying for full data access.

Requests for data access are reviewed by the Data Access Committee. During review, the primary investigator and submission preparer (if different from primary investigator) may be contacted for verification. The information you will need to gather to submit a data access request as well as a link to the form are listed below. For additional questions regarding data access, contact: [email protected]

Full data files are stored separately from this repo within Azure. To gain access to these data files, any researcher may submit a data access request, which includes:

• Principal investigator information
  • Academic credentials, affiliation, contact information, curriculum vitae, signature attesting accuracy of data use request
• Additional investigator information
  • Academic credentials, affiliation, contact information
• Research proposal
• Acknowledgement to comply with Data Use Agreement. Key points are listed below:
  • No sharing of data with anyone outside of approved PI and other specified investigators. New investigators must be reviewed.
  • No data use outside of stated proposal scope
  • No joining of data with other data sources
  • No attempt to identify participants, contact participants, or reconstruct PII
  • Storage with appropriate access control and best practices
  • You may publish (or present papers or articles) on your results from using the data provided that no confidential information of Microsoft and no Personal Information are included in any such publication or presentation
  • Any publication or presentation resulting from use of the data should include reference to the Aurora-BP Study, with full reference to the source publication when appropriate
  • Aurora-BP Study authors and Microsoft are under no obligation to provide any support or additional materials related to the use of these data
  • Aurora-BP Study authors and Microsoft are not liable for any losses, damages, or harms of any kind in connection to the use of these data
  • Aurora-BP Study authors and Microsoft are not responsible or liable for the accuracy, usefulness or availability of these data
  • Primary Investigator will provide a signature of attestation that they have read, understood, and accept the Data Use Agreement