SMART ENROLLMENT ASSISTANCE: A Predictive AI Chatbot for University Admission Insights and Data Visualization

Project conducted: 2024-2025 | Awarded 2nd Prize at DUE Student Scientific Research Conference

---

Introduction

This project was developed as part of the Student Scientific Research Conference 2024-2025 at the University of Economics – University of Danang (DUE). Our team proudly achieved Second Prize with this work.

Project Objective:
To build an AI-powered platform that integrates data visualization, personality assessment, and a smart chatbot to support students and parents in accessing university admission information accurately, efficiently, and in a personalized way.

Note:
Currently, only the paper PDF is uploaded in this repository for readers to understand the system concept and detailed implementation. The code is not publicly available at this stage for intellectual property protection and potential future commercialization.

---

Platform Functions Overview

The platform includes three main features:

1. Overview Dashboard

Provides an interactive data visualization dashboard for users to:

• Explore university admission data (cut-off scores, quotas, majors, schools, years, combinations).
• Filter and analyze data dynamically based on personal preferences.
• Identify trends, averages, and comparisons across majors and universities to support informed decision-making.

Example:

Treemap of admission scores by major and university

Average admission score statistics by admission method

Average admission score trends by university

Comparison of admission scores across majors in the most recent year

Quick recommendations for majors and universities by score range

2. MBTI Test Module

An integrated MBTI personality assessment module allows users to:

• Complete a 70-question MBTI test online.
• Discover their personality type and its characteristics.
• Receive personalized career and major recommendations aligned with their MBTI type, aiding better academic and career orientation.

Example:

MBTI test interface

Recommended majors based on MBTI personality types

3. AI Chatbot for University Admission

This is the core highlight of the project:

• Functionality:
  Enables users to ask free-form questions about:

  • University majors, cut-off scores, quotas, and admission methods (using structured data from a scraped dataset).
  • General information about DUE such as mission, training programs, and scholarships (using unstructured text data).

• Technical Implementation:

  • Uses Retrieval-Augmented Generation (RAG) for context-aware, accurate responses.
  • Embedding with intfloat/multilingual-e5-large-instruct for semantic search, stored in ChromaDB Vector Database.
  • Combines Function Calling for structured queries with RAG for knowledge-based queries.
  • Built with GPT-4o mini for natural language generation.

• Workflow Summary:

Chatbot system architecture for university admission counseling

1. User submits a question.
2. System classifies intent (admission data lookup, general info, or casual query).
3. For structured data queries (e.g. cut-off scores), it uses Pandas for precise table filtering.
4. For unstructured knowledge (e.g. program description, mission), it performs semantic search via embeddings + RAG to retrieve top-k documents as context for the LLM to generate final answers.
5. Outputs are formatted clearly for user reading.

Example:

Chatbot interface

Chatbot response result to user query

• Evaluation Results:
  • Faithfulness score: avg. 0.8996
  • Answer relevancy score: avg. 0.8717

Evaluation chart of chatbot answer faithfulness and relevancy

• Response time:
  • Normal queries: ~1–2s
  • Complex RAG queries: ~30–40s

Comparison of execution time across different question types

For detailed technical architecture, evaluation metrics, and module design, please refer to the attached paper.

---

How to Use

To experience the platform, visit:
SMART ENROLLMENT ASSISTANCE Demo

Website interface

---

Learning Outcomes

Through this project, we:

• Developed skills in web scraping, data engineering, and dashboard visualization.
• Integrated psychological assessment (MBTI) into educational recommendation systems.
• Built and deployed a practical AI chatbot with RAG + Function Calling for real-world applications.
• Enhanced teamwork, project management, and presentation abilities, culminating in an award-winning research project.

---

Documentation

• Project Paper PDF

---

Note on Code Availability

The code for this project is currently not public to protect intellectual property and due to potential future commercialization. For any collaboration or inquiry regarding this work, please contact the project team.