/davinci

Primary LanguageJupyter Notebook

DAVINCI: Visual Analytics Tool for Political Discourse Analysis

Pipeline Political discourse analysis plays a crucial role in understanding societal trends, shaping public opinion, and holding leaders accountable in democratic societies. However, the complexity and volume of political discourse data pose significant challenges for researchers and analysts.

We introduce DAVINCI, a visual analytics tool designed to facilitate the analysis of political discourse, with a focus on Mexican presidential speeches. DAVINCI employs advanced text preprocessing techniques, topic modeling algorithms, and intuitive visualization methods to uncover temporal trends, sentiment shifts, and associations with external events within political discourse data.

Through a multifaceted evaluation approach, including a detailed case study and qualitative user feedback, we demonstrate the effectiveness and usability of DAVINCI in extracting meaningful insights from political discourse data.

Accessing the Code

All code can be accessed in the root_directory/davinci directory. For example, if the root directory is /home/user/projects, then the code would be located at /home/user/projects/davinci.

An example of the visualization tool in HTML format

You can check out the live version here.

Installation Instructions

  1. Clone the repository:

    git clone https://github.com/dapivei/davinci.git
  2. Navigate to the project directory:

    cd davinci
  3. Install the package using pip:

    pip install -r requirements.txt.

Repository Structure

  • data/: Folder containing datasets used in the analysis.
    • (Content updated last month)
  • docs/: Documentation files for the project.
    • imgs/: Images used in the documentation.
      • (Content updated last month)
  • notebooks/: Jupyter notebooks for exploratory analysis and demonstrations.
    • (Content updated last month)
  • src/: Source code of the DAVINCI tool.
    • visutils.py: Utility functions for visualization.
    • (Content updated 3 days ago)
  • .gitignore: Specifies intentionally untracked files to ignore.
    • (Content updated last month)
  • README.md: Main documentation file.
    • (Content updated last month)
  • environment.yml: Environment configuration file.
    • (Content updated last month)
davinci/
│
├── docs/
│   ├── discourse_analysis_visualization.pdf
│
├── imgs/
│   ├── ...
│
├── notebooks/
│   ├── ...
│
├── src/
│   ├── visutils.py
│   ├── ...
│
├── data/ (optional)
│   ├── clean/
│   ├── prepro/
│   └── raw/
│
├── .gitignore
├── README.md
└── environment.yml