Aviation Satisfaction Analysis

Air travel is one of the most data-rich industries in the world, yet passenger dissatisfaction remains stubbornly high. This project asks a simple but commercially important question: what actually makes a passenger leave happy?

Working with a dataset of 103,594 passenger logs, I built an end-to-end analysis pipeline in Python. Starting with data cleaning and feature engineering, moving through exploratory data analysis, and finishing with a Random Forest classification model that predicts passenger satisfaction with 96% accuracy. Along the way I engineered new features like Age Group segmentation, Total Delay, and an Average Service Score composite to give the model richer patterns to learn from. Statistical significance testing using Chi-Square and Point-Biserial correlations ensured that every insight presented was mathematically grounded, not just visually convincing.

The findings challenged some common assumptions. Delays, while frustrating, don't lose customers on their own. Passengers with moderate delays were still satisfied when the rest of their experience held up. Food and drink, despite being a common airline talking point, showed far weaker predictive power than expected. The real dealbreakers were digital: inflight Wi-Fi and online boarding quality were the top two drivers of satisfaction by a significant margin, even among Economy passengers.

To make these findings accessible beyond a technical audience, I built an executive summary dashboard in Tableau Public. It features interactive service impact toggles, an Age Group by Class heatmap, a delay scatter plot, and a service rating gap chart that ranks where satisfied passengers consistently scored higher. The dashboard is designed so that a non-technical stakeholder can draw the same conclusions as the data scientist who built the model.

Tools used: Python (Pandas, Scikit-learn, Matplotlib, Seaborn), Tableau Public

Key skills demonstrated: Feature engineering, classification modeling, cross-validation, statistical testing, dashboard design, business storytelling