OPINION MINING ON TESLA (2015) USING TWITTER DATA

Abstract

In the current technological landscape, the surge in social media[1] usage has become a predominant platform for the exchange of ideas and beliefs. Opinion mining also referred to as sentiment analysis, has emerged as a pivotal tool for understanding public sentiment. Through the application of Natural Language Processing (NLP), sentiment analysis auto- mates the extraction of attitudes, opinions, and emotions from diverse sources, including text, audio, tweets, and databases. Leveraging data from Kaggle, a renowned hub for data science and machine learning, our research focused on analyzing tweets pertaining to top companies from 2015 to 2020 to gain comprehensive insights into international public sentiment trends. Twitter, with its widespread accessibility, facilitated swift and efficient data collection, offering valuable insights into ongoing topical discussions. Our study concentrated specifically on examining approximately 30,000 tweets about Tesla from early 2015. Employing methodologies such as CountVectorizer and various classifi- cation algorithms including Naive Bayes, Decision Trees, Random Forests (RFC), Logistic Regression, XGBoost, and Support Vector Machines (SVM), we sought to categorize and assess the emotions elicited by these tweets. Our findings underscored the efficacy of Ran- dom Forest and SVM in providing the highest classification accuracy, thereby significantly contributing to a nuanced understanding of public sentiment dynamics surrounding Tesla and, by extension, other top companies. This research not only sheds light on the evolv- ing landscape of public opinion but also underscores the potential of sentiment analysis techniques in informing decision-making processes across various domains.