Smart Classroom Analytics: Adaptive Learning Pathways using Eye-Tracking and Interaction Data for Computer Education

Project Abstract

This study presents a novel framework for enhancing computer education through smart classroom analytics that integrates eye-tracking, interaction data, and adaptive learning pathways. The core aim is to design an unobtrusive, real-time analytics system that observes student engagement, cognitive load, and learning progress to dynamically tailor instructional content and pacing. By combining eye-tracking metrics (fixation duration, saccade patterns, pupil dilation) with interaction data (mouse/keyboard events, code execution, error rates, and time-on-task), the framework infers affective and cognitive states such as confusion, frustration, and flow, enabling precise diagnostics of learning bottlenecks. A multi-layer architecture is proposed a sensing layer that collects high-frequency biometric and behavioral signals, a processing layer that fuses multimodal data using machine learning models (including sequence models and graph-based representations) to detect learning states, and an adaptation layer that delivers individualized learning pathways, resources, and remediation prompts through an intelligent tutoring system interface. The methodology includes data collection from diverse computer science cohorts across introductory and intermediate topics, followed by robust labeling through expert-annotated checkpoints and student self-reports to train supervised models. Key contributions include (1) a validated sensor fusion pipeline that preserves privacy and ensures user comfort, (2) interpretable models that map observable signals to cognitive and affective states with high accuracy, (3) a dynamic content adaptation engine that personalizes problem sets, hints, and coding tasks in real time, and (4) an evaluation framework combining quantitative learning gains, engagement metrics, and qualitative feedback. The study employs a quasi-experimental design with control, baseline, and intervention groups across multiple modules such as fundamentals of programming, data structures, and software development practices. Outcomes measured include learning gains (pre/post assessments, concept inventories), time-on-task, completion rates, error trends, and qualitative perceptions of usefulness and ease of use. Expected results indicate that students in the adaptive condition exhibit higher retention of concepts, more efficient problem solving, and increased sustained attention, while maintaining privacy and minimizing disruption to the learning process. Sensitivity analyses examine robustness to sensor noise, variability in individual learning styles, and cross-cultural differences in educational settings. The research also explores ethical considerations, including consent, data ownership, and transparent explanations of how analytics influence instructional decisions. Finally, the study discusses scalability implications for large classrooms, potential integration with existing learning management systems, and pathways for extending the framework to collaborative and project-based computer science curricula. The proposed system aspires to bridge the gap between educational data analytics and pedagogically meaningful interventions, delivering measurable improvements in learning outcomes while supporting educators with actionable insights and streamlined monitoring capabilities.

Project Overview

What This Project Is About

---

A straightforward exploration of how classroom analytics can tailor learning in computer education. The project looks at how students interact with computer-based lessons and where their attention goes, to create personalized learning paths that adapt to individual needs.

---

---

The Problem It Addresses

---

Many online and blended courses use the same content for all students, which can slow some and overwhelm others. Wasteful time on material that isn’t challenging or engaging reduces learning gains. This project seeks practical ways to adjust teaching materials in real time based on student behavior.

---

---

Objectives of the Project

---

1. Identify cues of student engagement and confusion from eye-tracking and interaction data.
2. Develop a simple framework to adapt learning paths based on these cues.
3. Evaluate the usefulness of adaptive paths for improving learning outcomes in introductory computer courses.
4. Provide guidance for teachers on interpreting analytics and adjusting instruction.

---

---

What You Will Do Step by Step

---

1) Review basic concepts of educational analytics and accessibility considerations.

2) Collect data from volunteer participants using a lightweight eye-tracking setup and built-in activity logs.

3) Label moments of focus, distraction, and difficulty from the data, with simple criteria.

4) Design a small adaptive module that changes the next learning activity based on detected signals.

5) Test the module in a controlled classroom-like setting and gather feedback.

---

---

Expected Outcome

An accessible prototype showing how real-time data can guide personalized learning paths, plus insights on what types of adaptive changes help students most in computer education.