LLM-Driven Learning Analytics Dashboard for Teachers in EFL Writing Education

TL;DR

This paper introduces a human-centered LLM-powered dashboard that helps EFL teachers analyze student interactions with ChatGPT in writing education, improving monitoring and instructional strategies.

Contribution

It presents a novel dashboard integrating NLP and HCI principles to enhance teacher monitoring of ChatGPT-assisted EFL writing education.

Findings

Effective monitoring of student behavior with ChatGPT

Identification of noneducational interactions

Enhanced instructional strategy alignment

Abstract

This paper presents the development of a dashboard designed specifically for teachers in English as a Foreign Language (EFL) writing education. Leveraging LLMs, the dashboard facilitates the analysis of student interactions with an essay writing system, which integrates ChatGPT for real-time feedback. The dashboard aids teachers in monitoring student behavior, identifying noneducational interaction with ChatGPT, and aligning instructional strategies with learning objectives. By combining insights from NLP and Human-Computer Interaction (HCI), this study demonstrates how a human-centered approach can enhance the effectiveness of teacher dashboards, particularly in ChatGPT-integrated learning.