Sentiment Analysis for Brand Reputation: A Comparative Study of Technology Firms

This thesis carries out a comparative study of four brands in the technology sector, namely Apple, Samsung, Google and Microsoft, with the use of sentiment analysis to evaluate brand reputation. It is evaluated based on the data collected from three online platforms (Reddit, Trustpilot and Technology related news sources). The study employs a multi-method approach by integrating lexicon-based, machine learning and transformer-based sentiment analysis. In addition, the study uses topic modeling to explore specific topics and themes prevailing in user-generated content. It also applies aspect-based sentiment analysis to capture insights related to drivers of reputation such as product quality, innovation, customer service, usability, privacy and sustainability. The study computes key performance indicators such as share of voice, engagement rate, sentiment polarity and net brand reputation %, to evaluate the brand reputation. The study contributes to literature by combining multiple analysis techniques and applying them across diverse platforms.