QF5214_Group 8

Group Project of Course QF5214

Members: Gao XuanRong, Hur Sinhaeng, Li LingYan, Liu Yang, Ren ZhiNan, Zhang YiChen, Zhou Zheng, Zhang Leyan, Lee Jiazhe, Mei Su

Project Title

"From Sentiment to Strategy: Integrating NLP-Derived Factors into Quantitative Investment Models"

Project Objective

To extract market sentiment from textual data, quantify it into sentiment factors, and integrate these factors into a multi-factor investment model. We will build a real-time, interactive dashboard that visualizes all relevant data dynamically.

Streamlit Dashboard App

This repository hosts a Streamlit application deployed from StreamlitAPP/Home.py, designed to serve as a dynamic and interactive dashboard.

🌐 Live App: [The page has been removed. Kindly refer to the video recording of the page in QF5214 Project Presentation File or QF5214 GRoup 8 Dashboard Video file]

---

🚀 Navigating the Codebase [The SQL Cloud Data server has been closed]

Welcome! Below is a brief guide to help you explore our project repository with ease.

⚠️ Update
🛑 The SQL Cloud Data server is no longer maintained. If you need to inspect or work with the data, please refer to the .csv files provided in the Team1/, Team2/, and Team3/ directories.
These CSVs reflect the original data structure and are safe to use without affecting the integrated pipeline.

⚠️ Kind Reminder
🛑 Kindly do not run the scripts in Team1/, Team2/, or Team3/ folders, as doing so would previously modify the data stored in the SQL server, directly impacting the entire data pipeline and final integrated output in Team4/.
Now, with the server shut down, those scripts are retained for reference only. Please only execute from the Team4/ directory, which contains the finalized and fully integrated pipeline.

⚠️ 温馨提醒
🛑 SQL 数据服务器现已关闭。如需查看数据，请使用 Team1/、Team2/ 和 Team3/ 文件夹下提供的 .csv 文件，这些文件已完整保留原始数据结构以供参考。
请勿运行 Team1/、Team2/ 或 Team3/ 文件夹中的脚本，这些脚本曾用于修改 SQL 数据库，可能会干扰最终在 Team4/ 中整合的数据流程。
当前这些脚本仅供参考，实际操作请以 Team4/ 文件夹为准。

🗄️ Database Info

Update: The SQL Cloud Data server is no longer available. Please refer to the provided database diagram for all related information

If you need to inspect the data directly, please connect using the following database settings: - DB_NAME: QF5214 - DB_USER: postgres - DB_PASSWORD: qf5214G8 - DB_HOST: pgm-t4n365kyk1sye1l7eo.pgsql.singapore.rds.aliyuncs.com - DB_PORT: 5555

This diagram outlines:

• Data Structure: across different team folders and SQL schema
• Pipeline Timeline: including preprocessing, transformation, and integration
• Our Hybrid Processing Model: which combines:
  • Scheduled triggers: (e.g., hourly/daily batch jobs for enrichment & aggregation)
  • Streaming output: (e.g., near real-time updates pushed to downstream systems)

This architecture enables modular development by team, while preserving global consistency in the final integrated result.

📁 Code Structure Overview

We’ve structured the repository to reflect our collaborative stages:

• Team1/: Data Extraction & Preprocessing
• Team2/: Sentiment Analysis
• Team3/: Factor Modeling, Backtesting
• Team4/: Integration, visualization,and Dashboard

Each folder contains a detailed README to further explain the scope, methodology, and key deliverables of that stage. These READMEs provide:

• A brief summary of the team's objective
• A description of the data used and generated
• Key scripts and their functions
• Instructions for running the code
• Dependencies and environment setup
• Any assumptions or limitations
• This modular structure ensures clarity, facilitates team collaboration, and allows for independent development and testing of each component.