Data and code for yet another adventure in machine learning for Ars Technica.
The data here is from the UC Irvine Machine Learning Repository, originally derived from the work of medical research at the Cleveland Clinic and the Hungarian Institute of Cardiology. (See the UC Irvine page for cites.) I went back to this source after looking at data on Kaggle derived from the original set without citation; it's in public domain.
cleve-hung-fixed is the combination of the Cleveland Clinic and Hungarian Institute data sets.
A data analysis Jupyter notebook is included.