Global goals, local voices: A multinational comparative sentiment and topic analysis of public transportation in the context of SDGs

Abstract

In the contemporary world, public transportation emerges as a fundamental component of societal infrastructure and becomes a critical driver of economic growth, environmental sustainability, and social equity. The increasing demand for public transportation, population growth, and urbanization underscores the need for a comprehensive understanding of public sentiment toward these services. Such insights are crucial for governments to align service offerings with public expectations and optimize resource allocation. This thesis delves into the exploration and analysis of public opinion on public transportation in five culturally and geographically diverse countries, using data collected from Twitter. By applying advanced techniques in natural language processing, specifically sentiment analysis and topic modeling, the study aims to examine public perception, identifying areas of satisfaction and dissatisfaction. This analysis not only promises to offer valuable insights for enhancing transportation services but also aims to contribute to the broader discourse on sustainable development, aligning with the United Nations' Sustainable Development Goals. The use of social media data in this research underscores the potential of digital platforms as rich sources of public opinion, offering a novel approach through which urban planning and policy-making can be informed and enriched. This thesis builds upon key theories and models of sentiment analysis, topic modeling, and public transportation research. A significant research gap identified is the application of NLP techniques, to understand public opinions on public transportation systems. The thesis aims to bridge this gap by analyzing public sentiment and issues related to public transportation, thereby contributing to the understanding of public perception and its alignment with the Sustainable Development Goals. This approach challenges the current scope of sentiment analysis and topic modeling applications by extending them into the domain of public transportation. The study employs a mixed-methods approach, combining quantitative data analysis with qualitative insights. It focuses on sentiment analysis and topic modeling of tweets to determine public opinions about public transportation in the UK, USA, India, South Africa, and Australia. Data collection involved gathering 406,005 tweets via the Twitter API from January 1, 2021, to March 31, 2023, focusing on the terms 'public transportation' and 'public transport.' These tweets, characterized by their dates, user locations, and text content, underwent preprocessing, including cleaning and normalization steps like removing URLs, emojis, and stop words and applying stemming and lemmatization. The study then utilized the RoBERTa language model for sentiment analysis and Latent Dirichlet Allocation for topic modeling. The outcome of this analysis was then paired with relevant Sustainable Development Goals (SDG) targets to assess progress and perceptions in five different countries towards sustainable public transportation. The key findings reveal significant public concerns about transportation costs, infrastructure quality, safety, and service quality across the studied nations: the UK, USA, India, South Africa, and Australia. Notably, in the UK and Australia, mandatory use of masks during the COVID-19 pandemic received mixed reactions, indicating varying public perceptions of health safety measures in public transport. The study also uncovered specific challenges like women's safety in public transport in India and South Africa, highlighting the need for targeted interventions. Furthermore, the adoption of electric buses in India reflects a significant shift towards sustainable transportation. These findings have profound implications for policymakers and stakeholders in the transportation sector. They underscore the need for inclusive, equitable, and sustainable transportation policies that address public concerns and align with Sustainable Development Goals, particularly in ensuring safety, affordability, and accessibility in public transportation. The thesis offers pivotal insights into public perceptions of public transportation, uncovering the diverse impacts of geographic, cultural, and policy factors. Key findings reveal significant regional variations in sentiment and priorities, with safety, infrastructure, and costs emerging as universal concerns. The study's novel approach in pairing public sentiment with specific Sustainable Development Goals highlights the role of public transportation in achieving broader societal objectives, including gender equality, education, and public health. This work significantly contributes to the literature by providing a comprehensive, data-driven understanding of public transportation issues across various contexts, enhancing our grasp of how different factors play in shaping public opinion. The research lays the groundwork for future explorations into a broader, multi-lingual analysis of public transportation, extending to other service sectors and evolving user behaviors.