A COMPARATIVE SAAS FRAMEWORK FOR REAL-TIME SOCIAL MEDIA SENTIMENT ANALYSIS USING MULTI-MODEL APPROACH

Papiya Mukherjee; Parul Saini; Priyanka Tyagi; Shweta Singh; Varsha Srivastava; Mamunur Islam

doi:10.29121/shodhkosh.v7.i10s.2026.8187

Authors

Papiya Mukherjee School of Computer Science & Engineering, IILM University, Greater Noida, India
Parul Saini School of Computer Science & Engineering, IILM University, Greater Noida, India
Priyanka Tyagi School of Computer Science & Engineering, IILM University, Greater Noida, India
Shweta Singh School of Computer Science & Engineering, IILM University, Greater Noida, India
Varsha Srivastava School of Computer Science & Engineering, IILM University, Greater Noida, India
Mamunur Islam School of Computer Science & Engineering, IILM University, Greater Noida, India

DOI:

https://doi.org/10.29121/shodhkosh.v7.i10s.2026.8187

Keywords:

Sentiment Analysis, Llm, Supervised Machine Learning, Openai, Web Scraping, Data Dashboarding

Abstract [English]

The process of examining a piece of text to identify whether the underlying sentiment expressed is positive, negative, or neutral is known as sentiment analysis. While it may seem straightforward, sentiment analysis is currently playing a vital role in comprehending how people perceive and interact on social media platforms, which is critical information for businesses and content creators. By gauging the sentiments conveyed through textual data, companies and individuals creating online content can gain valuable insights into the emotional responses and engagement levels of their target audiences. This paper presents our platform “Analytix” which is designed for analyzing sentiments across YouTube, Facebook, Reddit, Twitter and WhatsApp etc. It supports 197 languages and can handle multiple languages at once. The core of our platform is its user-friendly dashboard that pulls out the sentiments (positive, negative, or neutral) from posts across different social media platforms. It offers users the flexibility to choose from three different model options for sentiment analysis: GPT, transformer models, and a custom proprietary sentiment analysis model developed inhouse. It deeply analyzes each post to identify the overall sentiment people have towards it

References

“API Reference.” Streamlit. https://docs.streamlit.io/develop/ api-reference (accessed April 10, 2024).

“Documentation” gpt-4-turbo-and-gpt-4. https://platform.openai.com/ docs/models/gpt-4-turbo-and-gpt-4 (accessed March 30).

“Reddit API Documentation” Reddit. https://docs.streamlit.io/develop/ api-reference (accessed date April 15, 2024).

“Transformers” Hugging face. https://huggingface.co/transformers/v3.5. 1/quicktour.html (accessed April 7, 2024).

“Youtube Data API” References. https://developers.google.com/youtube/ v3/docs/search/list (accessed April 5, 2024).

Alqahtani, A., Ilyas, H., Naeem, M. W., and Bashir, S. (2023). “A Comprehensive Survey on Multilingual Sentiment Analysis”, IEEE Access, vol. 11, pp. 30048–30071. IEEE.

Anukarsh, G. P., Sanjana, S., Skanda M. B., and Harish, B. S. (2017). “Sentiment Analysis for Sarcasm Detection on Streaming Short Text Data”, 2nd International Conference on Knowledge Engineering and Applications, IEEE, 2017.

Biswas, E., Karmaker Santu, S. K., and Chakraborty, S. (2023). “Ex-amining the Impact of ChatGPT and LLMs on Sentiment Analysis and Opinion Mining”, Proceedings of the 2023 Conference on Empirical Methods in Natural Language Processing (EMNLP), pp. 3214–3229. ACL.

Bonta, V., Kumaresh, N., and Janardhan, N. (2022). “A Comprehensive Study on Lexicon Based Approaches for Sentiment Analysis”, Asian Journal of Computer Science and Technology, vol. 8, no. S2, pp. 1–6.

Chen, Z., Zhang, Y., and Liu, B. (2022). “Sentiment Analysis in the Era of Large Language Models: A Reality Check”, arX iv preprint arXiv:2305.15005, 2023.

Liu, Y., Ott, M., Goyal, N., Du, J., Joshi, M., Chen, D., Levy, O., Lewis, M., Zettlemoyer, L., and Stoyanov, V. (2020). “RoBERTa: A Robustly Optimized BERT Pretraining Approach”, Proceedings of the International Conference on Learning Representations (ICLR 2020).

Mondher, B., and Tomoaki, O. (2017). Apatternbased approach for sarcasm detection on twitter. IEEE. Transl. 2017

Naseem, U., Razzak, I., Musial, K., and Imran, M. (2020). “Transformer Based Deep Intelligent Contextual Embedding for Twitter Sentiment Analysis”, Future Generation Computer Systems, vol. 113, pp. 58–69. Elsevier.

Neri, F., Crisci, A., Agresta, S., Ambrosini, L., Cucchiarelli, A., & Stramaglia, S. (2018). Analyzing Sentiments in Social Media: A Case Study on News Broadcasts. In Proc. of the 9th International Joint Conference on Knowledge Discovery, Knowledge Engineering and Knowledge Management (IC3K 2017) (pp. 419-428). SCITEPRESS.

Tadesse, Y. (2026). Ai-Powered Security Strategies for The Osi Model. ShodhAI: Journal of Artificial Intelligence, 3(1), 9–19. https://doi.org/10.29121/shodhai.v3.i1.2026.62

Yadav, A., and Vishwakarma, D. K. (2020). “Sentiment Analysis Using Deep Learning Architectures: A Review”, Artificial Intelligence Review, vol. 53, no. 6, pp. 4335–4385. Springer.