Predictive Analysis of Customer Booking Completion with British Airways: Harnessing Data Mining Techniques

In this study, advanced data mining techniques were applied to predict and categorize customer ticket sales for British Airways based on online behaviour. Leveraging a unique dataset comprising 50,000 observations and 14 characteristics provided by British Airways, sophisticated data preprocessing, including the management of outliers, character variables, and missing data, was undertaken. Two machine learning models Random Forest and Decision Trees were employed and evaluated for their performance. The Random Forest model exhibited remarkable accuracy, achieving 93%, underscoring its effectiveness in forecasting customer booking patterns. In contrast, the Decision Tree model yielded a moderate accuracy of 63.5%, with potential for improvement through hyperparameter tuning. These outcomes contribute to the ongoing discourse on refining customer booking predictions, showcasing the Random Forest model’s prowess and identifying areas for enhancement in the Decision Tree model. The study concludes with actionable recommendations for British Airways, emphasizing timely customer engagement, flexible flight schedules, tailored packages, online platform improvements, enhanced in-flight experiences, group booking accommodations, optimized pricing strategies, and catering to dietary preferences.