This GitHub repository contains the materials related to a thesis project aimed at utilizing sentiment analysis to examine tweets concerning COVID-19 and related topics. The primary objective is to gain insights into how the population has responded to these themes and how their way of life evolved from the beginning of 2021 to the end of 2022.
In this thesis, we delve into the analysis of sentiment by employing advanced techniques to explore the sentiments expressed within tweets during the critical period of the global COVID-19 crisis. By using sentiment analysis algorithms, we dissect the data collected from Twitter to identify prevailing emotions, opinions, and trends among users. The thesis aims to identify key discussion topics, such as the dissemination of news, safety measures, vaccination campaigns, and the social and economic impacts of the pandemic.
Our approach involves working with an extensive dataset of tweets gathered over the designated timeframe. We process this dataset using state-of-the-art sentiment analysis algorithms to classify sentiments as positive, negative, or neutral. Moreover, we go beyond this classification to individually identify 28 distinct emotions, including anger, happiness, sadness, fear, surprise, disgust, and more. This granularity allows us to deeply understand the population's emotional spectrum.
Temporal changes are also considered, enabling us to discern any shifts in the expression of emotions and sentiments over time. By combining sentiment analysis with temporal analysis, we extract nuanced insights into the evolving emotional landscape.
The outcomes of this research furnish a comprehensive overview of the public's reaction to COVID-19 related themes. The results illustrate how emotions and opinions have transformed over the course of the examined period. This deeper comprehension of public sentiment facilitates a better understanding of the evolving public opinion and how the pandemic has influenced people's way of life.
This repository contains the following materials:
- Code: The codebase for the sentiment analysis, including data preprocessing, sentiment classification algorithms, and temporal analysis scripts.
The insights obtained from this study are invaluable for informing future crisis management strategies and public communication efforts. By understanding how sentiment evolves in response to global health crises, policymakers, healthcare professionals, and communicators can tailor their strategies to effectively manage similar situations in the future.
Feel free to explore the repository and engage with the materials. Your feedback and contributions are greatly appreciated.
For inquiries, please contact Mattia Bellisai.
Disclaimer: This repository is for educational and research purposes only. The data used is anonymized and publicly available. The opinions and sentiments expressed in the tweets are those of the users and do not represent the views of the researchers or this repository.