Research Data and Code for “Automated Violence Assessment in Video Games Using Metadata Analytics and Audience Engagement”

Description

This dataset contains the data, code, result tables, diagnostic appendices, and figures supporting the manuscript “Automated Violence Assessment in Video Games Using Metadata Analytics and Audience Engagement.” The study uses publicly available video game metadata, ESRB ratings, and YouTube audience engagement indicators to validate a metadata-driven automated violence assessment framework. The package includes the final analysis dataset, seed game catalogue, sanitized Jupyter notebooks, Python code exports, regression outputs, diagnostic files, codebook, README, and manuscript figures.

Steps to reproduce

Download and unzip the research data and code package. Open data/games_analysis_final_v6.csv, which contains the final analysis dataset used in the manuscript. Review documentation/CODEBOOK.md or documentation/CODEBOOK.csv for variable definitions. Open the sanitized notebooks in the code/ folder. Run Games_Analysis_Pipeline_Final_Updated_sanitized.ipynb to reproduce the data preprocessing, variable construction, engagement index, regression models, and diagnostic outputs. Optional: run YouTube_Games_Analysis_V11_Hybrid_sanitized.ipynb to reproduce the YouTube metadata collection workflow. This step requires a valid YouTube Data API key; replace YOUR_YOUTUBE_API_KEY_HERE with your own key. Compare generated outputs with the files in the results/ folder, including OLS models, log-log models, HC3 robust standard errors, VIF diagnostics, Breusch-Pagan results, Cook’s distance, residuals, and correlation tables. Figures used in the manuscript are available in the figures/ folder.

Institutions

• Simon Fraser University

  British Columbia, Burnaby

Categories

Licence