Data-driven analysis of TCAS Engineering Programs using Flask, Pandas, and Typhoon LLM

TCAS Dashboard is a web application for analyzing and comparing per-semester tuition costs of Computer Engineering programs across universities in Thailand. The project combines Data Engineering, NLP, and LLM-assisted information extraction into a single, end-to-end pipeline.

This repository is intentionally designed from a senior / production-oriented perspective, focusing on:

• Clear data pipeline separation
• Real-world Thai unstructured text challenges
• Explicit comparison between rule-based (Regex) and LLM-based extraction

The raw data is collected from the MyTCAS API and processed using Typhoon AI (LLM) before being visualized through a clean and interactive Flask Dashboard.

• Automated Data Collection from MyTCAS API
• Text Cleaning & Normalization (Regex vs LLM)
• LLM-assisted Information Extraction using Typhoon AI
• Interactive Dashboard with tables, charts, and cost rankings
• Experimental-ready Architecture for extending to other engineering majors

Example visualizations: tables and charts ranked by per-semester tuition cost

This repository intentionally implements two different approaches for extracting tuition cost information from Thai text.

👉 Only the LLM-based results are used in the production dashboard, as they provide significantly better robustness and coverage for real-world data.

TCAS_dashboard/
│
├── app.py                 # Flask application for rendering the dashboard
├── scraping_typhoon.py    # Data pipeline: fetch → clean → extract using Typhoon LLM
│
├── data/                  # Cleaned datasets (.csv / .xlsx)
│   ├── regex_cleaned/
│   └── llm_cleaned/
│
├── scripts/               # Notebooks and scripts for scraping and preprocessing
│
├── templates/             # HTML templates and dashboard assets
│   ├── dashboard.html
│   └── *.jpeg
│
├── experimental/          # Experiments extending the same logic to other engineering fields
│                           # (e.g., Electrical, Civil Engineering)
│
├── scrap_regex.ipynb      # Baseline: Regex-only extraction (no LLM)
├── requirements.txt       # Python dependencies
├── .env.example           # Environment variable template
├── .gitignore
└── README.md

git clone https://github.com/xooooiz7/TCAS_dashboard.git
cd TCAS_dashboard

python3 -m venv .venv
source .venv/bin/activate

pip install -r requirements.txt

python app.py

Open your browser at: http://127.0.0.1:5000

To understand the limitations of rule-based text extraction:

jupyter notebook scrap_regex.ipynb

cp .env.example .env
# Add your API key
# TYPHOON_API_KEY=YOUR_TYPHOON_API_KEY

python scraping_typhoon.py

The cleaned output will be stored in the data/ directory and used by the dashboard.

• Clear separation between data ingestion, processing, and visualization layers
• Easily swappable data sources and extraction strategies
• LLM usage is scoped only to tasks where rule-based methods do not scale
• Repository structure is designed for extensibility, not just demo purposes

• Add filtering by university and region
• Compare tuition costs with curriculum quality proxies
• Cache LLM responses to reduce API costs
• Production deployment (Docker + Gunicorn)

This project is for educational and engineering experimentation purposes only. It is not an official system of TCAS or MyTCAS.

Built with ❤️ using Flask, Pandas, and Typhoon AI