Sarina Sugawara

An attempt to visualize and quantify the qualitative data.

This will serve as a preliminary project for the authors’ forthcoming JALT research project scheduled for fall 2025. The primary objective of this project is to conduct an experimental analysis of sentiments by examining the textual input generated by non-linguistic major students in Japan.

The author attempts to explore the application of text mining methods to unveil hidden sentiments in textual data. By employing visualization techniques using an open-source pretrained model, or a transformer library for machine learning created by Hugging Face Inc. 

The project aims to reveal covert propensities and emotional undertones, providing insights into the underlying motivations and attitudes reflected in the text written by the college students.

The findings from the pilot study indicated that textual entries containing the term ‘English’ or ‘英語’ indeed reveal negative feelings among students. 

Furthermore, if the emotional categories were more complex, a notable degree of neutral sentiments were linked to their choice of words.

Speaker Bio: Sarina Sugawara

Adjunct faculty at Akita Prefectural University and 2024 JALT Research Grant recipient.  Committed to enhancing instructional design and analysis in higher education, with a passion for CLIL-related field and Cross-Cultural Studies.

*Hugging Face is an open-source machine learning platform and community that allows users to collaborate, train, and deploy on models, datasets, and applications.  https://huggingface.co/

This is Ms. Sugawara’s first official presentation for Akita JALT as a member, so if you are interested, please come and listen.

The presentation will be on May 18th (Sunday) from 2 pm to 3 pm through ZOOM.

All Akita Jalt members will be emailed the invitation to join in May.

Anyone else interested can email for the Zoom information at: akita@jalt.org

Thank you very much, and hope to see you there!