This project aims to detect and analyze the sources cited in news articles related to conflicts 🚨, to determine if there's any bias based on the origin of these sources. By examining the balance of source origins, such as whether sources are predominantly US-based versus local, the tool provides insights into the impartiality of conflict coverage in news articles. This project leverages articles from the BBC archives 📚 and utilizes libraries like Bertopic, HuggingFace, OpenAI, and Streamlit to achieve its goal. Developed as part of a hackathon, this tool is designed to promote transparency 🕊️ and encourage a more nuanced understanding of conflict reporting.
- Source Origin Detection 🔎: Automatically detects the origins of sources cited in conflict-related news articles.
- Bias Analysis ⚖️: Analyzes the balance of source origins to identify potential biases in news coverage.
- User-Defined Queries 🧑💻: Allows users to specify a conflict and the relevant sides (e.g., US-based vs. local sources) to tailor the analysis to their interests.
- Article Recommendation 📰: Suggests articles where one side or the other has greater prominence, based on the user's query.
- BBC Archives 📚: A rich dataset of news articles covering a wide range of topics, including conflicts, since 2017.
- Bertopic 🤖: For topic modeling and classification of articles. We use this to identify articles associated with different conflicts.
- HuggingFace 😊: Utilized for its powerful NLP models and tools.
- OpenAI 🧠: Provides advanced AI models for text analysis and insight generation.
- Streamlit 💻: Enables the creation of an interactive web application for users to engage with the tool.
- Python 3.11 or higher 🐍
- Pipenv for managing project dependencies 📦
- Clone the repository:
git clone https://github.com/camorales197/source_analyzer.git
cd source_analyzer- Install dependencies using Pipenv:
pipenv install- Activate the Pipenv shell:
pipenv shell- Launch the Streamlit web application 🚀:
streamlit run app.py-
Open your web browser and go to the URL displayed in your terminal (typically
http://localhost:8501). -
Use the application by entering the conflict and sides you are interested in analyzing 🕵️♂️.
We welcome contributions from the community 💖. If you're interested in helping improve this project, please feel free to reach out and let's talk about it.
This project is licensed under the MIT License 📄.