We used R to improve and predict operational outcome of the reactors. Data given to us had no labels, so we had no context of each independent variable. Linear regression models and tuned parameters achieved an average of 77% accuracy in prediction, similar to the company's results.