Used ggplot2 and cowplot libraries to construct bar plot, stacked bar chart, histogram, mosaic plot, corrplot to understand the relationship between the features.
Imputing missing values, changing features to appropriate data types and ROSE package to balance the target variable using over and under sampling.
Designed xgboost (with tuning the hyperparameters using caret) and randomforest model from scratch and compared the result with confusion matrix and roc curve and extracted importance variable for each.