data_analysis_platform

Data analysis platforms for viewing multimodal sensor data developed in Dash.

Setup

1. Python >= 3.6 is required.

2. Anaconda or other similar virtual environment is recommended.

3. Install Python modules dependencies.

pip install -r requirements.txt

If the auto install failed, you may try to install pakcages manually, e.g. pip install rich

Workspace layout

To run this data analysis tool, your workspace layout should look like this:

Workspace/
├── data_analysis_platform/
│   ├── fish_cutting_analysis/
│   │   ├── assets
│   │   └── app.py
│   ├── data_preprocess.py
│   ├── README.md
│   └── requirements.txt
└── data/
    ├── trial1/
    │   ├── cam_1-color-image_raw
    │   ├── cam_2-rgb-image_rect_color
    │   ├── ...
    │   └── Trial1_Cut_Fillets_Move_Toss.mat
    ├── trial2/
    │   └── ...
    └── ...

Usage

Data Preprocess

Some of the data might require preprocessing such as naming correction and color correction. Simply run

python data_preprocess.py

and follow the prompts to execute the functions you want.

Fish Cutting Analysis

1. Navigate to fish_cutting_analysis folder and run

python app.py

2. A folder selection gui will pop up, select the data that contains the trial data, for example <path_to_data>/trial1. Make sure the mocap data *.mat is also in the corresponding folder.

3. Visit http://127.0.0.1:8050/ in your web browser.

4. Drag the slider will fast forward the frames. The slider value is currently in milliseconds. You can set the hammer time for the camera view or mocap to sync data.

TODO

• Find faster methods to display video frames
• Maybe more elegant way to select data folders
• 3D plot support