A Lightweight Rule-Based Persian-English Sentiment Analysis Workflow Using n8n Automation

Mostafa Mahi; Zohreh  Arabi; Nadia  Bayzidi

doi:10.48314/des.vi.83

نویسندگان

Mostafa Mahi * Department of Computer Engineering and Information Technology, Payame Noor University, Tehran, Iran
Zohreh Arabi Department of Computer Engineering, Technical and Vocational University (TVU), Tehran, Iran
Nadia Bayzidi Department of Computer Engineering and Information Technology, Payame Noor University, Tehran, Iran

https://doi.org/10.48314/des.vi.83

چکیده

This paper presents a lightweight, fully offline workflow for Persian–English sentiment analysis implemented using the open-source n8n automation platform. This work presents a modular, low-code workflow integrating rule-based techniques for Persian–English sentiment analysis.

The workflow is designed for scenarios where computational resources, internet connectivity, and ML expertise are limited, prioritizing transparency, zero operational cost, and ease of deployment over robustness to highly noisy real-world data.

The system is evaluated as a proof-of-concept on a controlled synthetic dataset of 200 Persian and English comments, achieving 100% language detection accuracy and 92.5% sentiment classification accuracy under idealized conditions, with execution time below 1.8 seconds and memory usage under 100 MB. These results validate the functional feasibility of the proposed workflow. Future work will evaluate the system on real-world social media datasets.