Police body-worn cameras (BWCs), once limited to evidence collection, have evolved into powerful digital tools for real-time analysis of police behavior and community interactions. OpenBWC is an interdisciplinary project that integrates social science frameworks (e.g., Systematic Social Observation (SSO), Social Interactionist Theory (SIT)) with cutting-edge artificial intelligence (AI) techniques—computer vision (CV), natural language processing (NLP), and machine learning (ML).
Our goal is to detect, recognize, classify, and analyze patterns in BWC footage and related public data sources, providing actionable recommendations that can inform police training, improve transparency, and enhance community trust.
This repository focuses on data collection and preprocessing scripts that support this broader research vision.
yt-video-scripts/
Scripts for extracting and organizing YouTube data (e.g., news websites, publicly available BWC footage).
This is the starting point for gathering raw datasets that can later be analyzed using AI + social science frameworks.
More folders and modules will be added later as the project expands (e.g., CV/NLP preprocessing, annotation pipelines, and ML model training).
- Python 3.10+
- pip or Poetry for dependency management
- (Optional) Create a virtual environment:
python -m venv ~/.venvs/openbwc source ~/.venvs/openbwc/bin/activate