Emirates customer reviews — sentiment & text analysis
The business question
Airlines spend millions on service improvements, but without systematically analysing customer voice data, they're guessing at the pain points. This project asked: what are passengers actually unhappy about, and how does sentiment vary by route and cabin class?
What I did
Scraped and processed thousands of customer reviews using Python (Pandas, NumPy). Applied NLP techniques including TF-IDF keyword extraction and VADER sentiment scoring to categorise feedback by theme — in-flight service, food quality, punctuality, and staff behaviour. Visualised sentiment distributions and theme frequency across cabin classes.
Key finding
Economy class reviews were 2.3x more likely to mention "staff attitude" as a negative theme compared to business class, despite identical crew — suggesting a service consistency gap tied to workload distribution rather than individual performance.