Predicting civil rights issues in the airline industry: a machine learning approach using consumer complaint data

Purpose This study aims to predict civil rights issues in the airline industry using machine learning (ML) techniques, including neural networks (NNs), support vector machines (SVMs) and gradient boosting (GB). Design/methodology/approach Using 2015–2020 data from the US Department of Transportation (DOT)’s Air Travel Consumer Reports, the study analyzes civil rights complaints related to discrimination based on race, ancestry, national origin, color, religion and sex. Input variables included operational metrics (e.g. on-time performance, cancellations, delays, mishandled baggage) and consumer complaints (e.g. oversales, customer service, refunds). The analysis focuses on five major US airlines: Alaska, American, Delta, Southwest and United. Findings Oversales complaints, disability-related complaints and delays in the national aviation system emerged as key predictors of civil rights issues. These findings suggest that even issues not directly related to civil rights can escalate into civil rights allegations. The results offer insights for airlines to proactively manage such risks. Research limitations/implications The study is limited to five major US airlines and may not capture all incidents due to the scope of DOT data. Practical implications Airlines can use these findings to develop strategies to reduce civil rights complaints and improve service. Social implications Identifying key factors can help airlines foster a more inclusive environment and reduce discrimination in air travel. Originality/value This is the first study to apply advanced ML techniques to predict civil rights complaints in the airline industry, offering a data-driven approach to understanding and addressing discrimination.