Tasklist.txt

EARTHQWAZE TASKLIST:

- Complete Pipeline (front to back integration)	
	>> Inputs are a set of start and end coordinates
		>> python end needs to return set of coordinates that correspond to path to be drawn
		>> needs to be put into json_io.py(static) so that flask can serve GET and POST calls to the front end
	>> Make image processing and data upload to image classifier hands-free (involves rewriting matlab script in python)
	

- Hide API keys
	>> Locally generate a json file that we access to pull in api keys
	>> Do NOT push this file to the github repo (We can share it via email rather than pushing and pulling)
	>> bro.... you pushed this file already :) -Justin

- Rename Files 
	>> Some of the descriptions are currently redundant
	>> Rename files to better describe their functions
	>> some are refactored so far(index.html and some of the javascript ones)
	>> python files do need to get cleaned

- Document Code
	>> *Pretty self-explanatory


- Improve Simulated Disaster Data
	>> Gaussian noise method is a very elementary approach. Not robust to a wide variety of natural-disasters
	>> Conduct research on ways of modelling road damage in image processing.
	>> Could conduct actual tests to see how our model works so far


- Retrain/Refine IBM Visual Recognition Model 
	>> Train model over larger datasets (e.g. 1000+ images per level) now that color-data is enabled
	>> Generate more levels for damage classification (finer binnings)

- Fix some images to "declutter" main folder
	>> put images in Images folder
		>> I would mess with the map and graph but I haven't had a chance to look at the notebooks to see if they were hardcoded titles so I won't touch it yet so we don't have multiple copies
	>> consolidate all jupyter notebooks together
	>> refactor jupyter notebooks into one big python serving BEAST to finish off the project