Last Update: 09 March, 2018
The Beiwe Research Platform collects high-density data from a variety of smartphone sensors including GPS, WiFi, Bluetooth, and accelerometer. To learn more about Beiwe, check out the Onnela Lab page, the paper introducing the platform, or the Beiwe wiki.
In order to help current and potential collaborators understand the structure and format of Beiwe data, we are making our personal data available to the public. The hope is that access to real data will allow researchers to (1) facilitate coding and debugging for ETL, data ingestion, and other parts of their pipeline before data is collected, (2) create functions to help inspect raw data, and (3) test new methods or functions on real data.
Due to size limitations, this repository does not contain the data. To get the data, please download it from Zenodo. This repository contains information about the data, and we will add example notebooks of manipulating the data here. The original, raw release will always remain the same at the Zenodo page.
The Digital Object Identifier of this dataset is 10.5281/zenodo.1188879. When using these data, please cite the dataset as
Kiang, Mathew, Lorme, Jeanette, & Onnela, Jukka-Pekka. (2018). Public Sample Beiwe Dataset (Version 0.1) [Data set]. Zenodo. http://doi.org/10.5281/zenodo.1188879
When referring to the Beiwe Research Platform or how these data were collected, please cite the JMIR-Mental Health paper as:
Torous J, Kiang MV, Lorme J, Onnela JP, New Tools for New Research in Psychiatry: A Scalable and Customizable Platform to Empower Data Driven Smartphone Research, JMIR Ment Health 2016;3(2):e16. URL: https://mental.jmir.org/2016/2/e16, DOI: 10.2196/mental.5165
This work (both code and data) is licensed under a Creative Commons Attribution Share-Alike 4.0 License (CC-BY-SA-4.0).
Development of the Beiwe Research Platform and data analysis pipeline was enabled by National Institute of Health Director’s New Innovator Award to Dr. JP Onnela (DP2MH103909).
The original data come from 3 people using 5 different study settings of various durations and starting/ending dates.
In the public release dataset, all audio files have been removed (but users that collected audio files will still have an audio_recordings folder). Further, some GPS data have been anonymized by creating a bounding box around sensitive areas. Any observation within the bounding polygon is replaced anonymized by using last observation carried forward and adding a small amount of Gaussian noise.
Researchers may specify different data collection parameters for every study. The Beiwe app collects accelerometer, Bluetooth, call, GPS, gyroscope, magnetometer, power state, proximity, text, and WiFi data whenever possible or specified. Not all sensors are available on all phones. Call, power state, and text data were collected when events occurred, which, in these test data, was relatively infrequently. See the official Beiwe documentation for more details.
This table shows the study settings for each of the public-use studies. All durations are in seconds. The sensors are accelerometer ("Accel."), GPS, gyroscope ("Gyro."), magnetometer ("Mag."), DeviceMotion ("DM"; iOS devices only), Bluetooth ("BT"), and WiFi. Bluetooth is parameterized as amount of time spent on (logging nearby Bluetooth devices) out of a total amount of time. WiFi is parameterized as the frequency at which WiFi networks should be logged. All other sensors are parameterized as amount of time on and off.
| Study ID | Users (n) | First file timestamp | Last file timestamp | Accel. (on/off) | GPS (on/off) | Gyro. (on/off) | Mag. (on/off) | DM (on/off) | BT (on/total) | WiFi (log freq) |
|---|---|---|---|---|---|---|---|---|---|---|
onnela_lab_gps_testing |
1 | 2016-01-26 19_00_00 |
2016-02-04 14_00_00 |
10/10 | 7140/60 | 60/600 | 60/600 | 60/600 | 60/300 | 300 |
onnela_lab_ios_test1 |
1 | 2016-06-03 19_00_00 |
2016-06-04 16_00_00 |
10/900 | 60/1200 | 60/1200 | 60/600 | 60/600 | 60/300 | 300 |
onnela_lab_ios_test2 |
2 | 2016-06-07 18_00_00 |
2016-06-08 00_00_00 |
10/10 | 60/600 | 60/600 | 60/600 | 60/600 | 60/300 | 300 |
onnela_lab_test1 |
1 | 2016-10-17 19_00_00 |
2017-02-13 13_00_00 |
10/1250 | 60/1200 | 60/1200 | 60/1200 | 60/1200 | 60/300 | 300 |
passive_data_high_sampling |
1 | 2016-05-30 08_00_00 |
2016-08-25 07_00_00 |
600/60 | 7140/60 | 60/600 | 60/600 | 60/600 | 60/300 | 10 |
## Code can be found in "./code/plotting_hourly_heatmap_study1.R"
knitr::include_graphics("./plots/hourly_heatmap.jpg")
Similarly, we can aggregate over time and show heatmap of all GPS points on a map of Chicago.
## Code can be found in "./code/plotting_chicago_data.R"
knitr::include_graphics("./plots/chicago_map.jpg")