Running an airline company is extremely challenging due to stringent government approval processes and involves huge initial investments. To assist both new and existing players in the airline industry to improve customer experiences, we conducted an analysis of the Airline Passenger Satisfaction Dataset. This dataset contains approximately 104k customer satisfaction ratings on various aspects of the flight, including food, seat comfort, legroom, onboard Wi-Fi, and more.
In this project, we aimed to gain insights into key questions about the data and provide a comprehensive analysis of airline customer satisfaction levels. The following are the main tasks we performed:
Hypothesis Testing: We conducted hypothesis tests such as u-test and ks-test to explore relationships and make inferences from the data. One of our key objectives was to make predictions while controlling for confounding variables.
Model Development: We developed and compared multiple models to classify passenger satisfaction levels based on all available data in the dataset. The models we implemented included Linear Regression with Lasso and Ridge, logistic regression, random forest, and ADA boost. This allowed us to evaluate different approaches and select the most suitable model for our analysis.
Collaboration and Insights: Throughout the project, we collaborated as a team of 2 persons to ensure a comprehensive and insightful analysis. By leveraging our collective expertise, we aimed to provide valuable insights into airline customer satisfaction, which can contribute to improving customer experiences for both new and existing airline companies.
Our analysis of the Airline Passenger Satisfaction Dataset provided valuable insights into customer satisfaction levels in the airline industry. By employing various statistical tests and developing predictive models, we identified factors that significantly impact customer satisfaction and accurately classified satisfaction levels based on available data.
We believe that our findings and recommendations have the potential to significantly improve customer experiences for both new and existing airline companies. We encourage you to explore the detailed analysis and leverage the provided code and insights for further research and application.