My final project for the course "How to Win a Data Science Competition at Coursera", celebrated in October 2018. The project is actually a competition hosted by Kaggle:
https://www.kaggle.com/c/competitive-data-science-final-project
The project is split in two Jupyter notebooks. final_project_EDA contains the data exploration, while final_project_modelling contains the tasks of feature engineering, model optimization and ensembling. The notebooks should be self-explanatory. There is a script called data_io.py that is used by the notebooks to read the data files from disk and one configuration file called settings.ini with the data file paths.
The notebooks can be executed on a Python 3.6 environment with the libraries described in the file requirements.txt. To reproduce the results you just need to copy the competition datasets under the datasets/ folder and run the notebooks cells.
Final submission can be found in submission.csv and final models are also serialized in pickle files.