Bank Customer Reviews: Assessing Experiences Through Sentiment Analysis Using R Programming

Understanding customers’ emotional responses is critical for banks as digital channels proliferate and user-generated reviews grow. This study analyses online customer reviews from bankbazaar.com for six Indian banks (Axis Bank, HDFC Bank, Bank of Baroda, Canara Bank, Punjab National Bank and ING Vysya Bank) to map sentiment polarity and emotion-level patterns across public- and private-sector institutions. Using the syuzhet package in R with the NRC Emotion Lexicon, reviews were pre-processed (tokenized, cleaned, stop words removed) and scored for both polarity (positive/negative) and discrete emotions (trust, joy, anticipation, surprise, anger, fear, sadness, disgust). Results show an overall predominance of positive sentiment, with Axis Bank, HDFC Bank and Bank of Baroda displaying the strongest positive profiles and lower negative-word counts; conversely, Canara Bank, Punjab National Bank and ING Vysya Bank register higher negative-emotion frequencies. The article demonstrates the methodological value of lexicon-based sentiment analysis as a scalable, objective and reproducible alternative to survey-based approaches and offers emotion-aware insights that banks and regulators can use for targeted service improvements, real-time monitoring and policy design. Implications for customer experience management and future research directions are discussed.

JEL Codes: G21, G32, D81