As part of my graduate analytics coursework, I developed a Neural Network model using AI Studio (formerly RapidMiner) to predict airline customer satisfaction based on behavioral and service-level data. This project helped me explore how machine learning can be used not just to automate, but to truly understand what drives the customer experience.
đź’» Project Overview
Using a dataset of over 10,000 airline customer records, I created a Neural Network model that aimed to predict whether a customer would be satisfied based on inputs like:
- In-flight service quality
- On-time arrival performance
- Seat comfort
- Online booking experience
- Baggage handling and more
AI Studio allowed me to rapidly prepare, train, and evaluate the model using a visual, no-code interface, while still applying the strategic thinking and data prep skills needed in traditional machine learning workflows.
🧪 My Process in AI Studio
- Data Prep & Exploration: Cleaned the dataset and selected relevant features based on correlation and business logic.
- Model Design: Chose a Neural Net due to its strength in capturing nonlinear patterns in behavioral data.
- Training & Tuning: The initial model achieved ~77% accuracy, with high performance in predicting satisfied customers.
- Testing on New Data: Applied the model to a new subset to evaluate generalization. Results were consistent, although confidence scores suggested that future testing would benefit from larger sample sizes.
🧠What I Learned
This project reminded me that model performance is only part of the story—interpretability, confidence intervals, and practical application are equally important. I also saw firsthand how tools like AI Studio can bridge the gap between technical modeling and business usability. It’s a platform that empowers teams to prototype quickly, iterate intelligently, and focus more on asking the right questions.
🛠️ Skills Used
- AI Studio (RapidMiner): Neural network model design, testing, and deployment
- Machine Learning: Model evaluation, confidence analysis, classification modeling
- Customer Analytics: Satisfaction prediction, feature selection, behavior-driven insights
- Business Application: Practical modeling with client-facing use in mind
👋 Let’s Connect
This is just one of many projects that reflect my passion for using data and AI to make smarter, more human decisions. Feel free to explore more of my work here, and let’s connect on LinkedIn!