This repository contains the research project "Policy Patterns," which analyzes U.S. congressional transcripts to understand the evolving responses to Mexican cartel drug trafficking. The study spans over two decades of data, highlighting shifts in policy discourse and sentiment.
The primary objective is to identify and analyze changes in the tone, themes, and policy approaches within congressional discussions over time, particularly regarding law enforcement, public health, international cooperation, and socioeconomic considerations.
The hypothesis under investigation is that the discourse in U.S. congressional sessions regarding Mexican cartel drug trafficking has broadened from a focus on law enforcement to a more integrated perspective, reflecting global drug policy trends.
- Data Collection: Transcripts from the official U.S. Congress website.
- Sentiment Analysis: Performed using the VADER tool from the NLTK library.
- Topic Modeling: Conducted via Latent Dirichlet Allocation (LDA) using the
gensimlibrary. - Visualization: Includes heatmaps and word clouds to illustrate key findings.
data/: Contains the datasets used for analysis.scripts/: Python scripts for data collection, preprocessing, analysis, and visualization.notebooks/: Jupyter notebooks with detailed analysis steps and outputs.outputs/: Generated visualizations and results for reference.docs/: Additional documentation and project reports.
The findings indicate a partial shift in discourse towards public health narratives, alongside a persisting emphasis on security-related discussions. The transition towards a fully integrated policy approach is still emerging.
The study reveals the nuanced progression of U.S. policy discussions on drug trafficking, suggesting ongoing developments towards a holistic policy framework.