This test was originally done on March 2017. The company liked it and I joined BW7 as Data Scientist. It is a bit embarasing for me to review the code some years later but it may be helpful to share it.
For executing it just run:
make run
from the root folder of this repository. This will generate a docker image and it will run it. From the shell one can copy the token for the Jupyter notebook session. Once within the session, just enter the notebooks folder and run Case study 1 and Case study 2.