DebArgVis

This is a lightweight tool for data visualization, currently focused on visualizing argumentative structure in debate. The tool consists of Python-based preprocessing and an interactive D3.js visualization, that can be viewed in the browser. In it's current version, DebArgVis can visualize Inference Anchoring Theory (IAT) annotated data.

Features

• Preprocessing
  • Data cleaning and normalization;
  • Chronological ordering;
  • Timestamp annotation;
  • Topic extraction.
• Visualization

Usage instructions

• Clone the repository
• Run the main.py file (visualizingDebates\main.py)
• Access the visualization in the browser via localhost (http://127.0.0.1:8000). The performance may vary depending on the browser, Firefox appears to be the best browser, Micrisoft Edge the worst
• After running the server once, the preprocessed data as well as the files for the visualisation can be accessed (and easily exported) via the analysis_results folder

Example visualization

License

MIT License

Acknowledgments

• D3.js
• Dataset QT30 (Hautli-Janisz et al. 2022)

Further notes

• Note that the video is not included in the repositiory, partially due to the file size
• As the data is not perfect, some connections/nodes cannot be displayed, the logger gives an overview of what had to be excluded