actipy

A Python package to process accelerometer data.

Axivity3 and Axivity6 (.cwa), Actigraph (.gt3x), GENEActiv (.bin), and Matrix (.bin) files are supported, as well as custom CSV files.

Axivity3 is the activity tracker watch used in the large-scale UK-Biobank accelerometer study.

🚀 Getting started

Prerequisite

• 🐍 Python 3.8 or greater

  $ python --version  # or python3 --version

• ☕ Java 8 (1.8.0) or greater

  $ java -version

📦 Install

$ pip install actipy

📖 Usage

import actipy

data, info = actipy.read_device("sample.cwa.gz",  # or "sample.gt3x.gz" or "sample.bin.gz"
                                 lowpass_hz=20,
                                 calibrate_gravity=True,
                                 detect_nonwear=True,
                                 resample_hz=50)

Output:

data [pandas.DataFrame]
                                 x         y         z  temperature      light
 time
 2014-05-07 13:29:50.430 -0.513990  0.070390  1.671922    20.000000  78.420235
 2014-05-07 13:29:50.440 -0.233940 -0.586568  0.082067    20.000000  78.420235
 2014-05-07 13:29:50.450 -0.080319 -0.950817 -0.810613    20.000000  78.420235
 2014-05-07 13:29:50.460 -0.067236 -0.975886 -0.865132    20.000000  78.420235
 2014-05-07 13:29:50.470 -0.109636 -0.857004 -0.508666    20.000000  78.420235
 ...                           ...       ...       ...          ...        ...

info [dict]
 Filename                 : data/sample.cwa.gz
 Filesize(MB)             : 69.4
 Device                   : Axivity
 DeviceID                 : 13110
 ReadErrors               : 0
 SampleRate               : 100.0
 ReadOK                   : 1
 StartTime                : 2014-05-07 13:29:50
 EndTime                  : 2014-05-13 09:50:33
 NumTicks                 : 51391800
 WearTime(days)           : 5.847725231481482
 NumInterrupts            : 1
 ResampleRate             : 50
 NumTicksAfterResample    : 25262174
 LowpassOK                : 1
 LowpassCutoff(Hz)        : 20.0
 CalibErrorBefore(mg)     : 82.95806873592024
 CalibErrorAfter(mg)      : 4.434966371604519
 CalibOK                  : 1
 NonwearTime(days)        : 0.0
 NumNonwearEpisodes       : 0
 ...

📚 Refer to the Glossary for a comprehensive list of outputs.

⚙️ Optional parameters

You can also specify time ranges and additional options:

data, info = actipy.read_device(
    "sample.cwa.gz",
    lowpass_hz=20,
    calibrate_gravity=True,
    detect_nonwear=True,
    resample_hz=50,
    start_time="2014-05-07 18:00:00",  # Start time (ISO format)
    end_time="2014-05-09 18:00:00",    # End time (ISO format)
    skipdays=1,                         # Skip first day
    cutdays=2,                          # Cut last 2 days
    start_first_complete_minute=True,  # Align to minute boundaries
    calibrate_gravity_kwargs={'stdtol_min': 0.01},  # Custom calibration params
    flag_nonwear_kwargs={'patience': '60m'}         # Custom nonwear params
)

📊 Processing a custom CSV file

You can also use the routines in actipy.processing to process custom CSV files, or for more fine-grained control:

import actipy.processing as P

data, info_lowpass = P.lowpass(data, 100, 20)
data, info_calib = P.calibrate_gravity(data)
data, info_nonwear = P.flag_nonwear(data)
data, info_resample = P.resample(data, 50)

See the documentation for more.

💻 Processing from the command line

We also provide a command-line tool to process a device file (currently, only AX3 and AX6 are supported) and output to CSV:

$ read_cwa sample.cwa.gz -o outputs --lowpass-hz 20 --resample-hz 50 --calibrate-gravity --detect-nonwear

Outputs:

• 📄 outputs/sample/sample.csv.gz
• 📋 outputs/sample/sample-Info.json

🤝 Contributing

If you would like to contribute to this repository, please check out CONTRIBUTING.md. We welcome contributions in the form of bug reports, feature requests, and pull requests.

📜 License

See LICENSE.md.