Simplifying Sentiment Analysis on Social Media: A Step-by-Step Approach
Xuan Truong Du Chau, Thanh Toan Nguyen, Jun Jo, Sara Quach, Liem Viet Ngo, Pham Thu Hien, Park Thaichon
Abstract
This tutorial presents a systematic guide to performing sentiment analysis on social media data, designed to be accessible to researchers and marketers with varying levels of data science expertise. We prioritise open science by providing comprehensive resources, including self-collected data, source code and guidelines, facilitating result reproduction. For marketing and business researchers without programming experience, this tutorial offers a robust resource for conducting sentiment analysis. Experienced data scientists can use it as a reference for evaluating cutting-edge approaches and streamlining the sentiment analysis process. Our work stands out in its unique perspective on the challenges and opportunities of sentiment analysis within the social media data domain. We delve into the potential of sentiment analysis for social media marketing, offering practical guidance and best practices for enhancing brand reputation and customer engagement. Notably, this tutorial advances beyond previous studies by comprehensively comparing a wide range of sentiment analysis methods, including state-of-the-art transfer learning approaches, filling a critical gap in the existing literature. Our commitment to transparency underscores our contribution, as we provide all necessary resources for result reproducibility. We make our resources available at the following address: https://tinyurl.com/SentimentTutorial .