Please check the pdf file to read the final report.

Customer satisfaction surveys are an important tool for airlines to improve their inflight experience and target customers. However, sometimes it is difficult for airlines to know which specific factors of a passenger’s inflight experience lead to a passenger’s satisfaction with the particular airline and which may deter the passenger from traveling with the airline again in the future. Therefore, the goal of this project is to predict whether a passenger will be satisfied or dissatisfied with their travel experience with a particular airline (name not given) based on various factors of the customer’s inflight experience, the customer’s purpose of travel, and characteristics of the customer.

Supervised learning methods will be used in order to predict passenger satisfaction. Models that will be used include, logistic regression, a decision tree model, a random forest model, a boosting model, and a neural networks model. Variable importance for each model will be used to determine which factors of a passenger’s experience are the most important determinants of satisfaction level. Additionally, clustering will be used to group passengers into segments. Such clusters have the potential to provide useful information to the airline about which group of customers are likely to have a satisfactory travel experience and which will have a negative travel experience. The airline could use information from these clusters to examine the strength and deficiencies in its inflight service for each market segment. For example if a particular cluster is dissatisfied with inflight entertainment, then the airline could target improving inflight entertainment for this market segment to improve customer satisfaction.