This project use scikitlearn to build four different decision tree models_two using entropy and two using gini impurity_to ascertain whether a potentially business-transforming deal with a mysterious coffee farm in China will take the business to the next level. Addtionally, a random forest was used for increasing predictive accuracy and controlling overfitting. RFmodel showed that 70% customers are likely to buy the coffee. It may be worth to strike the deal.