The code assumes the following project structure:
Project
|
+-- Data
| |
| +-- subject_name_1
| +-- subject_name_2
| |
| +-- correct
| |
| +-- subject_activity_correctness_timestamp.csv
| +-- subject_activity_correctness_timestamp.csv
| +-- incorrect
| |
| +-- activity_1.csv
| +-- activity_2.csv
|
+-- Notebooks
| |
| +-- 00_notebook.ipynb
| +-- 01_notebook.ipynb
|
+-- Plots
|
+-- utils (package)
| |
| +-- constants.py
| +-- file_readers.py
| +-- signal_processing.py
The data is organised by subject ID, and then by correctness. All files should be in csv format and the header is always 7 rows (recorded on the old Respeck V4 app).
The name of each recording file should follow the format: "subjectName_activityName_correctness_timestamp.csv"
The data is recorded at 25Hz and contains both accelerometer and gyroscope data. It is recorded using the Respeck V4 and should be downsampled to 12.5Hz before using. We do not use the gyroscope data in this project, but keep it in the files for future research.
Here is where all the Jupyter Notebooks live. Note that if you wish to import any code from the project packages,
you need to add these two lines at the top of every notebook:
import sys
sys.path.append(..)
Here is where all the utility functions exist. The most important file is file_readers.py, where all functions for reading and preprocessing signals live.