Table 2 Comparison of open-source wearables analytics packages
Package Name (Code) | Domain(s) | Modes (Sensor Types) | Inputs (Specific Device/Data) | Nodes(Body Site) | Additional Notes |
|---|---|---|---|---|---|
Single domain – single device | |||||
 SleepPy(Python) [58] | • Sleep | • Accel • Temp • Light | • GENEActiv | • Wrist | • While the primary focus is sleep, does provide activity index • Tables and charts for analysis and presentation of outputs and sleep measures for each day separately |
 GaitPy (Python) [59] | • Gait | • Accel | • Formatted CSV | • Lumbar | • Uses single axis (vertical) from low back site and requires subject height • Not currently maintained – Scikit Digital Health (listed below) includes a newer version of GaitPy |
 Verisense Toolbox Step Detection Algorithm (GGIR2.0) [60] | • Gait | • Accel | • Same criteria as GGIR | • Wrist | • Add-on to GGIR for step detection • Visualization of raw data • Summary of step count data |
 GENEAclassify (R) [61] | • Gait | • Accel | • GENEActiv | • Wrist | • Provides activity classification and intensity outcomes |
 Pampro (Python) [62] | • Activity | • Accel | • Axivity • GeneActiv • Actigraph • activPAL • Formatted CSV | • Wrist | • Visualization of raw data • Physical activity summary statistics • Extract bouts for specific cut points |
Single domain – multiple devices | |||||
 OpenSenseRT (Python and OpenSim) [63] | • Limb motion | Gyro Accel | • Raspberry Pi microcontroller • Adafruit IMU | • Many (up to 14) | • Limb kinematics (flexion, extension, rotation, adduction) • Visualizations and summary joint angles • Requires hardware/component assembly |
 BDlab-OR: FoGdetection (Matlab) [10] | • Gait | • Gyro • Accel | • Formatted CSV | • Lumbar • Bilateral feet or shanks | • Specific to detection of freezing of gait • Import bouts of data containing gait activity • Outputs number of freezing events |
 OpenIMU (Python) [64] | • Activity | • Accel • GPS | • Actigraph • OpenIMU logger • AppleWatch SensorLogger | • Wrist | • Visualization of time-series and summary activity metrics • Export data as CSV or Matlab format |
Multiple domains – single device | |||||
 GGIR (R) [65] | • Sleep • Activity • Sedentary | • Accel | • GENEActiv • Actigraph • Axivity • Movisens | • Wrist | • Visualization of data and tabular data across and within days • Verisense (noted earlier) is a GGIR add-on for step detection |
 Biobank Accelerometer Analysis | • Sleep • Activity • Sedentary | • Accel | • GENEActiv • Actigraph • Formatted CSV | • Wrist | • Visualizations and data outputs within and across days • Outputs time series of epoched data and bouts (CSV) • Originally designed to use with UK Biobank data |
Multiple domains – multiple devices | |||||
 Pfizer/Scikit-digital-health (Python) [70] | Sleep Gait Transition | • Accel | • GENEActiv • Axivity • APDM (Opal) | • Lumbar • Wrist | • Specific wear locations (nodes) used for specific domains run separately (wrist for sleep and activity; low back for gait and transitions) • Tables and charts for analysis and presentation (within and across days) • Signal processing utilities |
 NiMBalWear (Python) | • Sleep • Gait • Sedentary • Activity | • Accel • Gyro • Temp | • Axivity • GENEActiv • Actigraph • Bittium • Formatted EDF | • Wrist (uni- or bi-lateral • Shank (uni- or bi-lateral), • Trunk (sternum) | • Temporal synchronization of all devices (based on standard synchronization protocol) allowing for time series comparison and fusion • Modular design allows domains to be capture by different sensors (flexible sensor input set) • Tables and charts for analysis and presentation within and across days • Outputs resampled synchronized raw time series data (EDF) |