This paper is available on arxiv under CC 4.0 license.
Authors:
(1) Ajay Krishnan T. K., School of Digital Sciences;
(2) V. S. Anoop, School of Digital Sciences.
Table of Links
Climate change’s impact on human health poses unprecedented and diverse challenges. Unless proactive measures based on solid evidence are implemented, these threats will likely escalate and continue to endanger human well-being. The escalating advancements in information and communication technologies have facilitated the widespread availability and utilization of social media platforms. Individuals utilize platforms such as Twitter and Facebook to express their opinions, thoughts, and critiques on diverse subjects, encompassing the pressing issue of climate change. The proliferation of climate change-related content on social media necessitates comprehensive analysis to glean meaningful insights. This paper employs natural language processing (NLP) techniques to analyze climate change discourse and quantify the sentiment of climate change-related tweets. We use ClimateBERT, a pretrained model fine-tuned specifically for the climate change domain. The objective is to discern the sentiment individuals express and uncover patterns in public opinion concerning climate change. Analyzing tweet sentiments allows a deeper comprehension of public perceptions, concerns, and emotions about this critical global challenge. The findings from this experiment unearth valuable insights into public sentiment and the entities associated with climate change discourse. Policymakers, researchers, and organizations can leverage such analyses to understand public perceptions, identify influential actors, and devise informed strategies to address climate change challenges.
Keywords Climate change · Sentiment analysis · ClimateBERT · Public discourse · Natural language processing
1 Introduction
According to the World Health Organization (WHO), the greatest health threat to people in the twenty-first century is climate change. Between 2030 and 2050, this risk is expected to cause an additional 250,000 fatalities annually and express itself in various ways. Many of these health concerns can be decreased or avoided with prompt and effective adaptation, but doing so necessitates in-depth research and policies that are multi-sectoral, multi-system, and collaborative at several scales[Fard et al., 2022]. One of the most important issues that needs significant attention is climate change. The majority of scientists agree that human activity is accelerating the Earth’s climate change, which is having a disastrous effect on the world and its population. The consequences of climate change are more clear. In recent years, extreme weather occurrences like hurricanes, tornadoes, hail, lightning, fires, and floods have increased in frequency and intensity. As the world’s ecosystems change quickly, access to the natural resources and agricultural methods that support humanity is in danger. [Ardabili et al., 2020]. The problem of climate change is complicated, and there is no quick fix. But to identify solutions, it’s critical to comprehend the issue.
Large volumes of text data can be analyzed using natural language processing which may unearth interesting patternsAnoop [2023]. The natural language processing approaches can be applied to the climate change domain as well for finding the causes and leveraging patterns such as public sentiment and discourse towards this global issue. Recent years have witnessed many people using social media to share their views, concerns, and public opinions on any topic under the skyAnoop et al. [2023]Jickson et al. [2023]. This has caused a huge amount of unstructured but dynamic data to be generated in such platforms, which are goldmines for social science researchersJohn et al. [2023]Anoop and Sreelakshmi [2023]. Collecting, curating, and analyzing such data is crucial for finding public perceptions and viewpoints on socially relevant discussionsVarghese and Anoop [2022]Lekshmi and Anoop [2022]. Similarly, understanding public opinions and sentiments on climate change is crucial for public policymakers, governments, and other administrators to devise better policies and intervention measures to address the challenges.
In this research, natural language processing is used to examine tweets that discuss climate change. We use ClimateBERTWebersinke et al. [2021], a pre-trained language model trained on a large set of climate change-related documents, and fine-tune the same for sentiment classification tasks. The findings may be used better to understand the public’s understanding of climate change, and it can also be used to identify the key stakeholders in the climate change debate. The interesting insights from this project will provide a foundation for informed decision-making and policy formulation regarding climate change. Additionally, the findings will contribute to advancing NLP techniques and their application in climate change analysis.
1.1 Effects of Climate Change
The consequences of climate change are extensive and have a significant impact on many facets of our planet. Several important aspects that shed information on the effects of climate change have been highlighted in research articles. Climate change is already manifesting globally, with a notable increase in extreme weather events. The frequency and intensity of hurricanes, floods, and droughts have amplified, causing widespread destruction and loss of life. Coastal areas are also seriously threatened by increasing sea levels, which might result in massive population displacement, increased erosion, and flooding. As glaciers continue to melt, water supplies diminish, affecting regions dependent on glacial meltwater for agricultural, industrial, and domestic purposes. Climate change-induced shifts in environmental conditions are causing profound changes in plant and animal life. Species are forced to adapt or face extinction as they grapple with altered ecosystems and changing habitats. This disruption to biodiversity has cascading effects on ecosystem functioning and services, with implications for food security, ecosystem stability, and human well-being.
Another consequence of climate change is the heightened risk of diseases. As temperatures rise, disease-carrying organisms, such as mosquitoes, expand their geographic range, exposing previously unaffected regions to vector-borne illnesses. This poses a significant public health challenge, necessitating the development of effective strategies for disease prevention, control, and surveillance. The impacts of climate change extend beyond the natural environment, affecting societies and economies globally. Disruptions to ecosystems and weather patterns have severe social and economic repercussions, with vulnerable communities being disproportionately affected. Climate-induced events, such as extreme heatwaves, prolonged droughts, and intense storms, lead to the displacement of populations, loss of livelihoods, and increased socioeconomic inequality. Consequently, countries face significant challenges in managing the economic and social ramifications of climate change, including the need for adaptation measures and the transition to sustainable practices. In summary, climate change is causing a range of effects that reverberate across multiple dimensions of our planet. The effects are widespread and provide serious difficulties for human society and ecosystems, from extreme weather events and increasing sea levels to melting glaciers and biodiversity loss. Addressing climate change requires concerted global efforts to mitigate greenhouse gas emissions, enhance resilience, and foster sustainable development practices to safeguard the future of our planet and its inhabitants.
1.2 NLP for Climate Change Analysis
Natural Language Processing (NLP) is an emerging discipline in computer science that concentrates on the creation of algorithms and models to facilitate computers in comprehending, analyzing, and producing human language. In the context of climate change analysis, NLP techniques have proven to be invaluable in extracting meaningful insights from vast amounts of textual data, offering a new perspective on this critical global issue. One application of NLP in climate change analysis is the ability to analyze public opinion on the topic. By leveraging sentiment analysis techniques, researchers can gauge the prevailing sentiments, attitudes, and beliefs surrounding climate change. This understanding of public opinion is crucial for policymakers, as it helps them tailor communication strategies, design effective interventions, and foster public engagement in addressing climate change challenges.
Furthermore, NLP techniques allow for identifying key stakeholders involved in the climate change debate.Through the extraction and examination of written information from various outlets, including news articles, social media platforms, and scientific journals, scholars can discern the key individuals, organizations, and institutions influencing the conversation surrounding climate change. This knowledge provides valuable insights into the various perspectives, interests, and motivations shaping climate change discussions, facilitating informed decision-making and targeted engagement with relevant stakeholders. NLP also enables the tracking of the progress of climate change negotiations. By analyzing texts from international agreements, policy documents, and meeting transcripts, researchers can monitor the evolution of climate change discussions, assess the effectiveness of existing frameworks, and identify areas of convergence or divergence among different stakeholders. This monitoring capability helps policymakers and negotiators evaluate the efficacy of climate change policies, identify potential barriers to progress, and inform future negotiations and policy development.
Additionally, NLP techniques can be applied to monitor the impact of climate change on different regions of the world. By analyzing textual data from scientific reports, environmental assessments, and socio-economic surveys, researchers can gain insights into the specific vulnerabilities, risks, and adaptation strategies associated with climate change in different geographic areas. This information is crucial for policymakers and local communities to prioritize resources, implement targeted interventions, and build resilience against the impacts of climate change. In summary, NLP techniques offer a powerful toolkit for analyzing climate change-related textual data, enabling researchers to gain valuable insights into public opinion, identify key stakeholders, track the progress of climate change negotiations, and monitor the impact of climate change on different regions. By harnessing the potential of NLP, policymakers and researchers can enhance their understanding of climate change dynamics and develop evidence-based strategies for mitigation, adaptation, and effective decision-making in the face of this global challenge. The major contributions of this research may be summarized as follows:
• Conducts a detailed study on different approaches reported in the natural language processing literature on sentiment analysis using social media data.
• Explores the potential of using ClimateBERT - a pre-trained model on climate data, for the sentiment analysis of tweets on climate change.
• Conducts extensive experiments and reports the experimental comparisons with different machine learning algorithms on sentiment analysis using ClimateBERT
