A project where I utilized multiple datasets to build models (using random forest and XGBoost) that predicts which people are most likely to support a particular candidate in a political campaign. The data had graph-like features which I used to engineer new features.