Sentiment and Emotion Analysis of Climate Related Social Media Content

Abstract

This study proposes a domain-specific sentiment and emotion analysis framework to address the limitations of general tools in capturing nuanced environmental discourse on social media. Using a hybrid model combining Pointwise Mutual Information (PMI) and Logistic Regression, along with NRCLex for emotion detection and BERTopic for topic modeling, the research analyses social media posts from 2014 to 2023. The findings show a predominance of negative sentiment, particularly on Twitter and Facebook, while Instagram displays more positivity. Common topics include climate change, plastic pollution, and air quality. The proposed model demonstrates improved accuracy over existing tools and is validated through expert annotations, offering a robust, multi-platform approach to understanding public perceptions of environmental issues—beneficial for policymakers, researchers, and communicators.

Introduction

To analyze public sentiment and emotions related to environmental issues (e.g., climate change, pollution) across Twitter, Instagram, and Facebook from 2014 to 2023, using a hybrid sentiment analysis model based on Pointwise Mutual Information (PMI) and Logistic Regression, combined with emotion detection (NRCLex) and topic modeling (BERTopic).

Key Contributions

• Developed a domain-specific sentiment analysis framework that outperforms general tools like VADER, spaCy, and Senti.

• Conducted cross-platform, multi-year analysis (2014–2023) of social media data on environmental topics.

• Identified platform-specific sentiment trends and emotional expressions.

• Linked emotions to specific environmental themes using topic modeling.

• Included expert annotation for model validation and increased contextual accuracy.

Methodology

A. Data Collection

• Platforms: Twitter, Instagram, Facebook

• Sources: User comments, captions, and tweets using environmental keywords.

• Timeframe: 2014–2023

• Preprocessing: Cleaning, tokenization, stop-word removal, normalization.

B. Training Data

• Sentiment140 dataset: 1.6M tweets filtered to 1,804 environment-related tweets for classifier training.

C. Sentiment Analysis

• PMI-based model: Measures word associations with positive/negative sentiment via Semantic Orientation (SO).

D. Emotion Detection

• Used NRCLex to detect 8 emotions (e.g., fear, trust, anticipation) in negative sentiment posts.

E. Topic Modeling

• BERTopic identified key environmental themes:

  • Climate Change

  • Air Quality & Emissions

  • Recycling & Plastic

  • Biodiversity

  • Environmental Responsibility

F. Manual Annotation

• 100 tweets manually rated on a scale from -5 to +5.

• Annotators included 2 domain experts (with weighted input).

• Inter-Annotator Agreement (Cohen’s Kappa): 0.525 (moderate agreement)

G. Model Evaluation

• PMI model accuracy: 65%

• Compared with:

  • VADER: 64%

  • Senti: 57%

  • spaCy: 44%

• Expert human annotator (ecologist): 90% accuracy (upper benchmark)

Results & Insights

A. Sentiment Trends

• Twitter: Consistently negative, reflecting concern and criticism.

• Facebook: Moderately negative; less intense than Twitter.

• Instagram: Increasingly positive, supporting optimism or sustainability promotion.

B. Emotion Trends (in negative posts)

• Dominant emotions: fear, trust, anticipation

• Spikes in fear/trust during 2020 (possibly pandemic-related).

• Facebook showed more sadness and trust after 2020.

C. Topic-Specific Insights

• Climate change: Most dominant and widely discussed.

• Air quality: Focus on transportation, electric vehicles.

• Plastic/recycling: Individual vs corporate responsibility debates.

• Biodiversity: Concerns over species extinction (mostly on Twitter).

• Energy policy: Facebook users criticized political leaders.

D. Platform-Specific Engagement Patterns

• Instagram: Positive posts more likely to go viral (positivity bias).

• Twitter/Facebook: Negative posts attract more interaction (negativity bias).

E. Semantic Orientation of Words

• Positive: clean, planet, alternative, electric

• Negative: fire, oil, fossil, animal, pollution

• Neutral/mildly negative: gas, bag

Conclusion

This research demonstrates the value of sentiment analysis in uncovering public opinion on environmental issues over a ten-year span using data from multiple social media platforms. The study revealed a dominance of negative sentiment in environmental discussions, with climate change emerging as the most frequently mentioned topic. Other recurring themes included air quality, emissions, plastic waste, and recycling, indicating their continued public relevance. Emotion analysis showed that fear, trust, and anticipation were the most common emotional responses, reflecting the public’s complex emotional engagement with these issues. These findings offer valuable insights for policymakers and environmental advocates to shape more effective and resonant communication strategies. However, the study faced challenges in sentiment classification accuracy due to the prevalence of negative tone, and the use of sarcasm and irony in social media posts, which complicated accurate interpretation.

References

[1] Masson-Delmotte, P. Zhai, H. O. Pörtner, D. Roberts, J. Skea, P. R. Shukla, A. Pirani, W. Moufouma-Okia, C. Péan, R. Pidcock, S. Connors, J. B. R. Matthews, Y. Chen, X. Zhou, M. I. Gomis, E. Lonnoy, T. Maycock, M. Tignor, and T. Waterfield, ‘‘IPCC, 2018: Global Warming of 1.5 ?C. An IPCC Special Report on the impacts of global warming of 1.5 ?C above pre-industrial levels and related global greenhouse gas emission pathways, in the context of strengthening the global response to the threat of climate change, sustainable development, and efforts to eradicate poverty,’’ Intergovernmental Panel Climate Change (IPCC), Cambridge Univ. Press, Cambridge, U.K., New York, NY, USA, Geneva, Switzerland, p. 616, 2018, doi: 10.1017/9781009157940. [2] H. Dagevos and J. Voordouw, ‘‘Sustainability and meat consumption: Is reduction realistic?’’ Sustainability, Sci., Pract. Policy, vol. 9, no. 2, pp. 60–69, Oct. 2013. [3] O. Y. Adwan, M. Al-Tawil, A. Huneiti, R. Shahin, A. Abu Zayed, and R. Al-Dibsi, ‘‘Twitter sentiment analysis approaches: A survey,’’ Int. J. Emerg. Technol. Learn., vol. 15, no. 15, p. 79, Aug. 2020. [4] K. L. S. Kumar, J. Desai, and J. Majumdar, ‘‘Opinion mining and sentiment analysis on online customer review,’’ in Proc. IEEE Int. Conf. Comput. Intell. Comput. Res. (ICCIC), Dec. 2016, pp. 1–4. [5] Jumadi, D. S. Maylawati, B. Subaeki, and T. Ridwan, ‘‘Opinion mining on Twitter microblogging using support vector machine: Public opinion about state Islamic University of Bandung,’’ in Proc. 4th Int. Conf. Cyber IT Service Manage., Apr. 2016, pp. 1–6. [6] I. K. C. U. Perera and H. A. Caldera, ‘‘Aspect based opinion mining on restaurant reviews,’’ in Proc. 2nd IEEE Int. Conf. Comput. Intell. Appl. (ICCIA), Sep. 2017, pp. 542–546. [7] V. B. Raut and D. D. Londhe, ‘‘Opinion mining and summarization of hotel reviews,’’ in Proc. Int. Conf. Comput. Intell. Commun. Netw., Nov. 2014, pp. 556–559. [8] A. Jeyapriya and C. S. K. Selvi, ‘‘Extracting aspects and mining opinions in product reviews using supervised learning algorithm,’’ in Proc. 2nd Int. Conf. Electron. Commun. Syst. (ICECS), Feb. 2015, pp. 548–552. [9] M. Wöllmer, F. Weninger, T. Knaup, B. Schuller, C. Sun, K. Sagae, and L.-P. Morency, ‘‘YouTube movie reviews: Sentiment analysis in an audio- visual context,’’ IEEE Intell. Syst., vol. 28, no. 3, pp. 46–53, May 2013. [10] A.-M. Iddrisu, S. Mensah, F. Boafo, G. R. Yeluripati, and P. Kudjo, ‘‘A sentiment analysis framework to classify instances of sarcastic sentiments within the aviation sector,’’ Int. J. Inf. Manage. Data Insights, vol. 3, no. 2, Nov. 2023, Art. no. 100180. [11] H. Almerekhi, H. Kwak, and B. J. Jansen, ‘‘Investigating toxicity changes of cross-community redditors from 2 billion posts and comments,’’ PeerJ Comput. Sci., vol. 8, p. e1059, Aug. 2022. [12] E. Shamoi, A. Turdybay, P. Shamoi, I. Akhmetov, A. Jaxylykova, and A. Pak, ‘‘Sentiment analysis of vegan related tweets using mutual information for feature selection,’’ PeerJ Comput. Sci., vol. 8, p. e1149, Dec. 2022. [13] A. R. Pratama and F. M. Firmansyah, ‘‘COVID-19 mass media coverage in English and public reactions: A west-east comparison via Facebook posts,’’ PeerJ Comput. Sci., vol. 8, p. e1111, Sep. 2022. [14] M. O. Faruk, P. Devnath, S. Kar, E. A. Eshaa, and H. Naziat, ‘‘Perception and determinants of social networking sites (SNS) on spreading awareness and panic during the COVID-19 pandemic in Bangladesh,’’ Health Policy Open, vol. 3, Dec. 2022, Art. no. 100075. [15] A. Giachanou, I. Mele, and F. Crestani, ‘‘Explaining sentiment spikes in Twitter,’’ in Proc. 25th ACM Int. Conf. Inf. Knowl. Manage., Oct. 2016, pp. 2263–2268. [16] B. Sluban, J. Smailovic, M. Juric, I. Mozetic, and S. Battiston, ‘‘Com-munity sentiment on environmental topics in social networks,’’ in Proc. 10th Int. Conf. Signal-Image Technol. Internet-Based Syst., Nov. 2014, pp. 376–382. [17] E. Rosenberg, C. Tarazona, F. Mallor, H. Eivazi, D. Pastor-Escuredo, F. Fuso-Nerini, and R. Vinuesa, ‘‘Sentiment analysis on Twitter data towards climate action,’’ Results Eng., vol. 19, Sep. 2023, Art. no. 101287. [18] C. Cubukcu-Cerasi, Y. S. Balcioglu, A. Kilic, and F. Huseynov, ‘‘Embracing green choices: Sentiment analysis of sustainable consumption,’’Eurasia Proc. Sci. Technol. Eng. Math., vol. 23, pp. 253–261, Oct. 2023. [19] F. Jost, A. Dale, and S. Schwebel, ‘‘How positive is ‘change’ in climate change? A sentiment analysis,’’ Environ. Sci. Policy, vol. 96, pp. 27–36, Jun. 2019. [20] B. Dahal, S. A. P. Kumar, and Z. Li, ‘‘Topic modeling and sentiment analysis of global climate change tweets,’’ Social Netw. Anal. Mining, vol. 9, no. 1, pp. 1–20, Dec. 2019. [21] M. Lineman, Y. Do, J. Y. Kim, and G.-J. Joo, ‘‘Talking about climate change and global warming,’’ PLoS One, vol. 10, no. 9, Sep. 2015, Art. no. e0138996. [22] W. Shi, H. Fu, P. Wang, C. Chen, and J. Xiong, ‘‘Climatechange vs. globalwarming: Characterizing two competing climate discourses on Twitter with semantic network and temporal analyses,’’ Int. J. Environ. Res. Public Health, vol. 17, no. 3, p. 1062, Feb. 2020. [23] T. E. Taufek, N. F. M. Nor, A. Jaludin, S. Tiun, and L. K. Choy, ‘‘Public perceptions on climate change: A sentiment analysis approach,’’ GEMA Online J. Lang. Stud., vol. 21, no. 4, pp. 209–233, Nov. 2021. [24] L. Rocca, D. Giacomini, and P. Zola, ‘‘Environmental disclosure and sentiment analysis: State of the art and opportunities for public-sector organisations,’’ Meditari Accountancy Res., vol. 29, no. 3, pp. 617–646, Jun. 2021. [25] Y. Tao, F. Zhang, C. Shi, and Y. Chen, ‘‘Social media data-based sentiment analysis of tourists’ air quality perceptions,’’ Sustainability, vol. 11, no. 18,p. 5070, Sep. 2019.

Copyright

Copyright © 2025 Priyanka Bakki, G. Praveen Babu. This is an open access article distributed under the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.