Skip to content

An activity classification model based on self-supervised learning for wrist-worn accelerometer data.

License

Notifications You must be signed in to change notification settings

OxWearables/actinet

Repository files navigation

actinet

A tool to extract meaningful health information from large accelerometer datasets. The software generates time-series and summary metrics useful for answering key questions such as how much time is spent in sleep, sedentary behaviour, or doing physical activity. The backbone of this repository is the self-supervised learning of Hang et al.: https://www.nature.com/articles/s41746-024-01062-3

Install

Minimum requirements: Python>=3.9, Java 8 (1.8)

The following instructions make use of Anaconda to meet the minimum requirements:

  1. Download & install Miniconda (light-weight version of Anaconda).

  2. (Windows) Once installed, launch the Anaconda Prompt.

  3. Create a virtual environment:

    conda create -n actinet python=3.9 openjdk pip

    This creates a virtual environment called actinet with Python version 3.9, OpenJDK, and Pip.

  4. Activate the environment:

    conda activate actinet

    You should now see (actinet) written in front of your prompt.

  5. Install actinet:

    pip install actinet

You are all set! The next time that you want to use actinet, open the Anaconda Prompt and activate the environment (step 4). If you see (actinet) in front of your prompt, you are ready to go!

Usage

# Process an AX3 file
$ actinet sample.cwa

# Or an ActiGraph file
$ actinet sample.gt3x

# Or a GENEActiv file
$ actinet sample.bin

# Or a CSV file (see data format below)
$ actinet sample.csv

Troubleshooting

Some systems may face issues with Java when running the script. If this is your case, try fixing OpenJDK to version 8:

conda create -n actinet openjdk=8

Offline usage

To use this package offline, one must first download and install the relevant classifier file and model modules. This repository offers two ways of doing this.

Run the following code when you have internet access:

actinet --cache-classifier

Following this, the actinet classifier can be used as standard without internet access, without needing to specify the flags relating to the model repository.

Alternatively, you can download or git clone the ssl modules from the ssl-wearables repository.

In addition, you can donwload/prepare a custom classifier file.

Once this is downloaded to an appopriate location, you can run the actinet model using:

actinet sample.cwa -c /path/to/classifier.joblib.lzma -m /path/to/ssl-wearables

Output files

By default, output files will be stored in a folder named after the input file, outputs/{filename}/, created in the current working directory. You can change the output path with the -o flag:

$ actinet sample.cwa -o /path/to/some/folder/

<Output summary written to: /path/to/some/folder/sample-outputSummary.json>
<Time series output written to: /path/to/some/folder/sample-timeSeries.csv.gz>

The following output files are created:

  • Info.json Summary info, as shown above.
  • timeSeries.csv Raw time-series of activity levels

See Data Dictionary for the list of output variables.

Plotting activity profiles

To plot the activity profiles, you can use the -p flag:

$ actinet sample.cwa -p
<Output plot written to: data/sample-timeSeries-plot.png>

Crude vs. Adjusted Estimates

Adjusted estimates are provided that account for missing data. Missing values in the time-series are imputed with the mean of the same timepoint of other available days. For adjusted totals and daily statistics, 24h multiples are needed and will be imputed if necessary. Estimates will be NaN where data is still missing after imputation.

Processing CSV files

If a CSV file is provided, it must have the following header: time, x, y, z.

Example:

time,x,y,z
2013-10-21 10:00:08.000,-0.078923,0.396706,0.917759
2013-10-21 10:00:08.010,-0.094370,0.381479,0.933580
2013-10-21 10:00:08.020,-0.094370,0.366252,0.901938
2013-10-21 10:00:08.030,-0.078923,0.411933,0.901938
...

Processing multiple files

Windows

To process multiple files you can create a text file in Notepad which includes one line for each file you wish to process, as shown below for file1.cwa, file2.cwa, and file2.cwa.

Example text file commands.txt:

actinet file1.cwa &
actinet file2.cwa &
actinet file3.cwa 
:END

Once this file is created, run cmd < commands.txt from the terminal.

Linux

Create a file command.sh with:

actinet file1.cwa
actinet file2.cwa
actinet file3.cwa

Then, run bash command.sh from the terminal.

Collating outputs

A utility script is provided to collate outputs from multiple runs:

actinet-collate-outputs outputs/

This will collate all *-Info.json files found in outputs/ and generate a CSV file.

Citing our work

When using this tool, please consider citing the works listed in CITATION.md.

Licence

See LICENSE.md.

Acknowledgements

We would like to thank all our code contributors, manuscript co-authors, and research participants for their help in making this work possible.