🎓 Cổng Đại học – The Universities' Gateway

🌐 Visit Website • 💻 Tech Stack • 🗂 Database • 📊 Data Pipeline • 🧠 Client Logic • 🚀 Future

---

🌍 Overview

Cổng Đại học (“Universities’ Gate”) is a web platform that helps Vietnamese students explore university admissions, compare exams, and discover suitable majors.

It integrates a large-scale admissions dataset, smart score conversions, and a modern interface — making it a one-stop portal for college admission guidance in Vietnam.

✨ Key Features

1. 🧭 Career Personality Assessment — Take the Holland Test and discover your personality type.
2. 🧮 Admission Score Calculation — Combine GPA, exam results (TSA, HSA, V-ACT, SPT, H-SCA…), and international certificates (IELTS, SAT, etc.).
3. 🎯 University & Major Lookup — Explore suitable universities and majors based on your score.
4. 📊 Exam Distribution Visualization — Compare and analyze exam score distributions.
5. 🔄 Cross-Exam Score Conversion — Convert scores using linear or percentile-based models.
6. 📚 Additional Information Lookup — Access documents and policies from the Ministry of Education and Training.

---

💪 Strengths

✅ Simple, user-friendly interface
📊 Huge, well-structured data warehouse
🧠 Versatile and diverse features supporting many use cases

---

🧰 Tech Stack

Layer	Technologies
Frontend
Database
Data Processing

---

🗂 Database Architecture

The system stores admission and exam data in Supabase, structured as follows:

erDiagram
    school {
        text id PK
        text name
        text short_name
        int8 public
        text region
    }

    score {
        text id PK
        text school_id FK
        text method_id FK
        text subject_group_id FK
        int8 year
        float8 converted_score
        float8 score
        text major_id
        text major_name
        text note
    }

    score_industry {
        text score_id PK, FK
        text industry_l1_id FK
        text industry_l2_id FK
        text industry_l3_id FK
    }

    industry_l1 {
        text id PK
        text name
    }

    industry_l2 {
        text id PK
        text industry_l1_id FK
        text name
    }

    industry_l3 {
        text id PK
        text industry_l2_id FK
        text industry_l1_id FK
        text name
    }

    method {
        text id PK
        text name
    }

    subject_group {
        text id PK
        text subject1 FK
        text subject2 FK
        text subject3 FK
    }

    subject {
        text id PK
        text name
    }

    exam_distribution {
        text subject_id FK
        float8 min
        float8 max
        jsonb distribution
        text method_id FK
        int8 year
    }

industry_l1 ||--|{ industry_l2 : contains
industry_l2 ||--|{ industry_l3 : contains
score_industry ||--|| industry_l1 : participates
score_industry ||--|| score : participates
score }|--|| school : participates
score }|--|| method : participates
score }|--|| subject_group : participates
subject_group }|--|{ subject : contains
exam_distribution }|--|| subject : contains
exam_distribution }|--|| method : contains

Loading

📂 Explore database and data samples: 👉 Google Drive Folder

📊 Data Processing Pipeline

crawling_score.py – Crawl raw admission data from educational websites.

handle_converted.py – Linear score normalization.

get_converted_score.js – Percentile-based score conversion.

generate_score_industry.py – Build relationships between score and industry_l3.

Push to database – Upload cleaned and verified data to Supabase.

⚠️ Note

get_converted_score.js uses web modules for fast conversion but may introduce encoding inconsistencies → all data is finalized using Pandas.

Percentile-based conversion aligns equivalent percentiles between two exams.

The average margin of conversion error is ±2 points (on a 30-point scale) — acceptable for comparison.

🧮 Conversion Flow

Standardize method names: uttt, xtkh → thpt, thhb, dgsg, dghn, dgtd.

Apply linear conversion → unify score scales (thpt: 30, dgsg: 1200, etc.).

Apply percentile conversion → align all results on a 30-point scale.

🧠 Client-Side Architecture UX Enhancements File Description loading.js Displays loading screen toast.js Handles notification toasts Shared Models File Purpose code_model.js Stores user data html_code_consts.js HTML templates for UI components filter_model.js Encodes/decodes admission score results holland_model.js Encodes/decodes Holland test results query_model.js Stores query state and statistics score_convert_model.js Numeric helpers (rounding, conversion, sorting, averaging) Page-Specific Scripts HTML Page Script File Description compare.html compare.js Exam comparison and visualization convert.html convert.js Score conversion filter.html filter.js Admission filtering & major recommendation general_statistics.html general_statistics.js General stats display holland.html holland.js Holland personality test introduce_industry_school.html introduce.js Industry & school introduction page 🧾 Data Views View Description view_score Combines essential data for major recommendation view_admission_unit Joins score and industry data view_top_mean Ranks universities by median score (excl. GPA-based) view_top_mean_by_industry3 Ranks majors by median score view_specific_ratio Measures school specialization level 📚 References

https://diemthi.tuyensinh247.com/ – Main data source

Official publications from the Ministry of Education and Training (Vietnam)

Personal research and experience

🚀 Future Development

🤖 AI-based major recommendation (using semantic similarity)

📈 Predictive modeling for next-year cutoffs

🌍 English and international version for overseas students

🧩 Public API for education and research data access