README.md

Heatmap

Graphically view input from mat.

Setup

Built in Python3 and relies on the SciPy stack. On Windows, the easiest way to use the SciPy stack is through Anaconda.

1. Install the SciPy stack using your preferred method.
2. Install virtualenv (it may already be installed) with python -m pip install virtualenv
3. Install virtualenv to the env/ folder using python -m virtualenv env
4. Activate the env using env\Scripts\activate (Windows) or env/bin/activate otherwise.
5. With the env activated, install requirements with pip install -r requirements.txt

Run the software

1. Modify the configuration in serial_com/ for the size of the matt.
2. Use the Arduino IDE to install the sketch to the device.
3. Modify the configuration in heatmap/__init__.py for rows and columns and any other config.
4. Activate the env using env\Scripts\activate (Windows) or env/bin/activate otherwise.
5. Run configurations:
   1. To run with random sample data, run python run.py fake
   2. To get data from serial, run python run.py serial
   3. To save the data after you quit, add a filename to the command, i.e. python run.py serial push_up to save the session to data/push_up_TIMESTAMP.json'
      • While running, press Enter to save a new section in the data.
   4. To replay saved data, run python run.py replay FILENAME.json
      • you can also request a single frame by adding arguments, i.e. python run.py replay FILANAME 3 22 will grab index 22 from section 3 of FILENAME.

Analysis

1. Install TensorFlow: https://www.tensorflow.org/install/
2. To convert saved data into digestible formats, run python model.py create
   • results can be previewed as images by running python model.py preview
3. To train the model, run python model.py train
   • To train on top of a prior model, run python model.py train r