EMS-Prediction

EMS Motion Data Visual Analytics

######################################################################################## ######################################################################################## ########################################################################################

USAGE INFORMATION:

Versions: python v3.6.3, D3 v5.9.2, safari v12.1, chrome v73

Report: VAML_Project_Report.pdf

1. Download /dashboard-d3 entire directory ONLY (this includes the data in it)
2. In terminal navigate into /dashboard-d3 directory
3. Execute: python3 -m http.server
4. Open either safari or google chrome and clear cache
5. In broser navigate to: http://0.0.0.0:8000/
6. Click and interacte with dashboard

######################################################################################## ######################################################################################## ########################################################################################

GitHub: https://github.com/dmscul11/EMS-Prediction

Box for Original Data: https://vanderbilt.app.box.com/folder/66875332205

Python Background Server for D3: python3 -m http.server http://0.0.0.0:8000/

Tinker for CNN: /home/scullydm/DoDHandsFree/ (use python3.6, pip3.6): nohup python3.6 combined_NN.py &

combined-d3:

• both squares D3 implementation

dashboard-d3:

• final squares and dashboard D3 implementation

tree-d3:

• Tree D3 implementation

squares-d3:

• Squares D3 implementation

stacks-d3:

• Squares and Tree combined D3 implementation

Main Code:

• ***read_processed_data.py = Main function to read in processed data and run ML random forest

• visualizations.py = Main function to read in ML output and create visualizations**************

• combined_NN.py = Main function to read in processed imputated data and run CNN

    --> input from combined-data
    
    --> Update params in main() function, lines 165 - 167
  

Parse and Combine Data Code (Run just once, UNLESS EVENTS TIMING HAS CHANGED OR MORE DATA ADDED):

• imputate_data_small.py (combine_data.py) = Main function to read in processed data and output imputation to small row size and wrap the rest --> input from processed-data --> Update params in main() function, lines 242-248
• imputate_data_large.py (combine_data.py) = Main function to read in processed data and output imputation to max row size --> input from processed-data --> Update params in main() function, lines 242-248
• parse_data_files.py = preprocess raw data and split data into individual csvs by event files info and data collection --> input from raw-data and event-files

Run Once Code:

• combine_data.py = Main function to read in processed data, combine and preprocess data, and run Neural Network
• handheatmap_script_NNData.py = generates csv files from openpose output
• analyze-watchdata-separate-events.py = generate basic plots of apple watch data
• limbs_randomforest.py = run random forest predictor on openpose data
• parse_data_files.py = preprocess and split data into csvs by procedure and data collection

Combined-Data:

• csv imputated data files of all same size and all data types by experiment, participant, event, trail #, time zeroed out
• input to combined_NN
• output from imputate_data.py combining data type
• generated from processed-data

Processed-Data:

• csv files of data time specific to experiment, participant, event, trial #, timestamp, data type
• input to imputate_data.py
• output from parse_data_files.py
• generated from raw-data

raw-data:

• csv files of data time specific to experiment, participant, and data type
• generated from SQL pulls (apple watch) or from Box (MYO) or AWS (video and open pose output)

Event-Files:

• procedure event details csv files for data for each experiment and participant