Architecture of a web-based system for analysing the tone and thematic structure of text data

Abstract

The article discusses approaches to building a web-based text data analysis system that combines message sentiment classification and thematic modeling within a single software solution. The aim of the study is to develop the architecture and software implementation of a system capable of automating the full text processing cycle: receiving input data, preprocessing, determining emotional tone, identifying key topics, saving results, and presenting them to the user via a web interface.

The methodological foundation consists of natural language processing methods, text vectorization using TF-IDF, sentiment classification using a Support Vector Classifier, and topic modeling based on Latent Dirichlet Allocation. 

A modular approach to design was used: the client-side, server API, machine learning module, authentication mechanisms, and data storage were separated. As a result of the research, a system architecture was proposed that ensures the integrated execution of two analytical tasks—determining the text’s sentiment and identifying thematic patterns in the message corpus. The paper describes the main system entities, the mechanism for forming an analysis request, the method of presenting results in JSON format, and the approach to service deployment, in which the base API is implemented in C#, and the analytical processing module is implemented as a separate service on FastAPI.

The scientific novelty of this work lies in the combination of classical machine learning models for text analysis with a distributed service architecture of a web application, which allows separating the computationally intensive natural language processing module from the system’s application logic and simplifies scaling, maintenance, and further expansion of functionality.

The practical significance of the results lies in the ability to use the proposed solution for monitoring public sentiment, analyzing user feedback, conducting market research, preprocessing text corpora, and generating analytical reports. The developed system can serve as a foundation for further improvement through the expansion of language support, the adaptation of models to domain-specific corpora, the enhancement of result interpretability, and the implementation of more advanced visualization tools.

References

1) Hankar M., Kasri M., Beni-Hssane A. (2025). A comprehensive overview of topic modeling: Techniques, applications and challenges. Neurocomputing. Vol. 628. Article 129638. DOI: 10.1016/j.neucom.2025.129638.

2) Blei D. M., Ng A. Y., Jordan M. I. Latent Dirichlet Allocation. (2003). Journal of Machine Learning Research. Vol. 3. P. 993-1022.

3) Lee D. D., Seung H. S. (1999). Learning the parts of objects by non-negative matrix factorization. Nature. Vol. 401. P. 788-791. DOI: 10.1038/44565.

4) Deerwester S., Dumais S. T., Furnas G. W., Landauer T. K., Harshman R. (1990). Indexing by latent semantic analysis. Journal of the American Society for Information Science. Vol. 41, No. 6. P. 391-407. DOI: 10.1002/(SICI)1097-4571(199009)41:6<391::AID-ASI1>3.0.CO;2-9.

5) Hayat F., Shatnawi S., Haig E. (2024). Comparative Analysis of Topic Modelling Approaches on Student Feedback. Proceedings of the 16th International Joint Conference on Knowledge Discovery, Knowledge Engineering and Knowledge Management - Volume 1: KDIR. P. 226-233. DOI: 10.5220/0012890400003838.

6) Devlin J., Chang M.-W., Lee K., Toutanova K. (2019). BERT: Pre-training of Deep Bidirectional Transformers for Language Understanding. Proceedings of NAACL-HLT 2019. Minneapolis. P. 4171-4186. DOI: 10.18653/v1/N19-1423.

7) Grootendorst M. (2022). BERTopic: Neural topic modeling with a class-based TF-IDF procedure. arXiv preprint arXiv:2203.05794. DOI: 10.48550/arXiv.2203.05794.

8) Cortes C., Vapnik V. (1995). Support-vector networks. Machine Learning. Vol. 20. P. 273-297. DOI: 10.1007/BF00994018.

9) Diri G., Obiorah P., Du H. (2025). Comparative Study of Sentiment Analysis Techniques: Traditional Machine Learning vs. Deep Learning Approaches. InSITE 2025. P. 23. DOI: 10.28945/5490.

10) Pedregosa F., Varoquaux G., Gramfort A. et al. (2011). Scikit-learn: Machine Learning in Python. Journal of Machine Learning Research. Vol. 12. P. 2825-2830.

11) Scikit-learn developers. sklearn.svm.SVC. Scikit-learn documentation. URL: https://scikit-learn.org/stable/modules/generated/sklearn.svm.SVC.html (accessed: 21.04.2026).

12) Bird S., Klein E., Loper E. (2009). Natural Language Processing with Python. O'Reilly Media. 504 p.

13) Scikit-learn developers. sklearn.feature_extraction.text.TfidfVectorizer. Scikit-learn documentation. URL: https://scikit-learn.org/stable/modules/generated/sklearn.feature_extraction.text.TfidfVectorizer.html (accessed: 07.05.2026).

14) Scikit-learn developers. sklearn.pipeline.Pipeline. Scikit-learn documentation. URL: https://scikit-learn.org/stable/modules/generated/sklearn.pipeline.Pipeline.html (accessed: 21.04.2026).

15) Scikit-learn developers. sklearn.decomposition.LatentDirichletAllocation. Scikit-learn documentation. URL: https://scikit-learn.org/stable/modules/generated/sklearn.decomposition.LatentDirichletAllocation.html (accessed: 21.04.2026).

16) Microsoft. Create web APIs with ASP.NET Core. Microsoft Learn documentation. URL: https://learn.microsoft.com/en-us/aspnet/core/web-api/?view=aspnetcore-10.0 (accessed: 22.04.2026).

17) FastAPI. FastAPI documentation. URL: https://fastapi.tiangolo.com/ (accessed: 22.04.2026).

18) Lewis J., Fowler M. (2014). Microservices: a definition of this new architectural term. URL: https://martinfowler.com/articles/microservices.html (accessed: 22.04.2026).