Assessing the Usefulness of News Sentiment for Real-Time Airline Stock Prediction

A project by Steven VanOmmeren examining the impact of news sentiment on airline stocks using advanced machine learning techniques and real-time data analysis.

Overview

This research project leverages the Global Database of Events, Language, and Tone (GDELT) to analyze how news sentiment affects airline stock volumes in near-real-time. The commercial airline industry is unique in that adverse events (crashes, incidents, etc.) are highly publicized and can dramatically impact public trust and stock prices.

Key Features

• Real-time Analysis: Predicts stock price changes in 15-minute increments during trading days
• Large-scale Data: Analyzes sentiment from 1.3 million news articles mentioning major U.S. airlines
• Comprehensive Coverage: Focuses on 7 major U.S. commercial airlines from January 2018 to May 2025

Research Objectives

1. Examine the impact of adverse news events on airline stock prices at the near-real-time level
2. Identify and analyze adverse news events using GDELT data
3. Predict real-time stock volumes and price changes more accurately than existing models
4. Demonstrate the economic value of GDELT for business monitoring applications

Data Sources

• GDELT (Global Database of Events, Language, and Tone): Real-time news sentiment data
• Stock Market Data: Real-time stock prices and volumes for major U.S. airlines

Prerequisites

• Python 3.12
• uv package manager (recommended) or pip

Key Findings

This research demonstrates:

• Superior Performance: Models achieve better accuracy than state-of-the-art approaches in predicting stock volumes
• Real-time Capability: System can process and predict based on news sentiment in near-real-time
• Economic Value: GDELT data provides actionable insights for business monitoring
• Scalability: Framework can be adapted to other industries beyond airlines

Academic Paper

The complete research findings are documented in a formal academic paper located in the Paper/ directory. The paper includes:

• Comprehensive literature review
• Detailed methodology
• Statistical analysis and results
• Economic implications and applications

Citation

If you use this work in your research, please cite:

@misc{vanommeren2025airline,
  title={Predicting Intraday Trading Volume with News Sentiment: An Analysis of U.S. Airline Stocks},
  author={VanOmmeren, Steven},
  year={2025},
  url={https://github.com/svanomm/sentiment-volume-forecasting}
}

License

This project is licensed under the MIT License, however we make no claim as to the licenses of the packages or underlying data relied upon in this work. See the LICENSE file for details.

Contact

For questions or collaboration opportunities, please contact Steven VanOmmeren.

---