Emotion as a Lens for Historical Understanding in Archaeogaming: Applying LLM-Assisted Annotation and Network Analysis

This repository contains the anonymized data, Python scripts, and reproducibility materials for the study “Emotion as a Lens for Historical Understanding in Archaeogaming: Applying LLM-Assisted Annotation and Network Analysis”.

The study examines how players’ emotional responses are associated with different forms of historical understanding in archaeology- and history-related digital games, focusing on Civilization VI, Assassin’s Creed: Origins, and the Tomb Raider series. The dataset includes processed annotation results for basic emotions, complex emotions, and historical-understanding dimensions, annotation validation data comparing human annotations with model outputs, derived correlation results, and the network data used for visualization.

The repository includes:

1. Processed annotation data for the three case-study games;
2. Merged analytical data used to generate descriptive figures;
3. Annotation validation data and the reproducible calculation of Supplementary Table S3;
4. Game-specific correlation result files;
5. A combined correlation network file used for network visualization;
6. Python scripts for reproducing the annotation reliability table, correlation analyses, and figures;
7. A README file explaining the file structure and reproduction steps.

Raw player-review texts are not included because they were collected from online platforms and may be subject to platform terms of service, copyright restrictions, and user privacy considerations. The data provided here are anonymized and processed research outputs intended to support transparency and reproducibility.

The scripts use relative paths and can be run after downloading the repository by installing the required Python packages listed in `requirements.txt`.