Check our the posts here: space2vec.com
The project behind the code is talked about in detail throughout the blog posts. But this is where the cool code stuff happens!
You can find the feature engineered CSV from the autoscan project (under the "Features" heading) site here: http://portal.nersc.gov/project/dessn/autoscan/
We have supplied requirements.txt file which you can use to setup the right environment. This was made for Python 3.6, so if you are getting errors about missing versions or something similar try removing anything after the "==" for that library in the requirements.txt and run again.
See /xgboost-baseline for code
We pickled the feature engineered data for our above model, you can find the data here: https://drive.google.com/open?id=1Pa4-imVbK7yfZuCX3mfF-mMae1eyhQqo