US Airline Sentiment Analysis using BERT

A study of sentiment analysis of US Airline dataset using BERT model from Hugging Face. This notebook trains a classifier with the features produced by BERT to classify sentiments into Positive, Negative and Neutral. A separate notebook consisting of exploratory data analysis on the feedbacks received by the customers to study the distribution of sentiments across different categories and also to analyse the reasons for the sentiments to derive useful insights about the airlines.