Data Preparation

We took dataset from Skytrax - Air Travel review website ( as an independent customer forum, which has become a leading review site for airline, airport and associated air travel traveller reviews. This website has no financial association or affiliation with any airline or airport featured.

The Skytrax name is associated with Quality Excellence throughout the world by the air transport industry, and is recognised for it’s Airline and Airport Star Rating, the World Airline Awards and Airport Awards

Guided Analytics

This workflow was constructed to understand the consumers’ feedbacks and issues about airline's service quality to improve customer satisfaction.​

In this worklow there are four different analysis levels:

1- NPS(Net Promoter Score) Analysis
2- Correlation Analysis between overall ratings and touchpoints
3- Word clouds based on positive and negative comments
4- Predicting sentiments to understand a comment was written by customer is negative or possitive. The model was built on the "02_Building_Model" workflow is used for this issue. ​

Model Building

Comments were collected from SKYTRAX which are about customer comments related with their satisfaction level of airline companies​, has been processed in the workflow called "01_Data_Preparation" and in this workflow the processed data is being used to build a sentiment prediction model.

Gradient Boosted Trees is being used to build powerful sentiment prediction. After training the model, it is saved to be used in the "Guided Analytics" part and worfklow of this use case.

Subscribe to Customer Experience and Sentiment Analysis