Summary

This repository, yaojingang/geo-citation-lab, provides a public resource for researching Generative Engine Optimization (GEO) and AI search. It addresses how AI search platforms like ChatGPT, Google AI Overview, and Perplexity trigger searches, select sources, and absorb citation content. The project offers reproducible data, analysis scripts, and a curated library of academic papers for researchers, engineers, and power users interested in understanding AI search mechanisms, citation influence, and potential manipulation risks.

How It Works

The core of the project comprises two main components: GEO experiment data reports and a curated paper collection. The experiment data section details studies involving 602 designed prompts across multiple AI search platforms. It includes data on search triggers, source selection, and citation absorption, featuring 72 citation-level features derived from over 21,000 search-layer rows and 23,000 citation-level rows. Analysis scripts are provided for feature extraction, statistical analysis, and influence reporting. The paper collection is organized thematically, consolidating research on GEO, AEO, AI search citation mechanisms, and manipulation risks.

Quick Start & Requirements

The repository is designed for direct execution of provided scripts.

• Setup: Clone the repository. Navigate to the 01-geo-experiment-data-report directory. Copy .env.example to .env and configure environment variables for API keys.
• Dependencies: Requires Python 3 and standard libraries. Specific analysis scripts (e.g., analyze_influence.py, build_self_contained_html.py) are located in 01-geo-experiment-data-report/03-pipeline and 01-geo-experiment-data-report/04-repet.
• Documentation: Quick overview available in QUICK_REPORT.md. Full reports in HTML, Markdown, and PDF formats are in 01-geo-experiment-data-report/04-repet/. Paper collection details are in 02-geo-aeo-ai-search-papers/README.md.
• Live Site: https://yaojingang.github.io/geo-citation-lab/

Highlighted Details

• Experimental data covers 602 prompts, 432 A/B/C/D layered experiments across 3 AI search platforms.
• Analyzed 72 citation-level features with a 76.44% citation capture success rate from 18,151 successfully fetched pages.
• Curated library contains 41 research papers categorized into 7 themes related to GEO, AEO, and AI search.
• Includes raw data CSVs (features_all_platforms_72.csv) and analysis pipeline scripts for reproducibility.

Maintenance & Community

The repository is hosted on GitHub (yaojingang/geo-citation-lab). No specific community channels (e.g., Discord, Slack) or details on active maintenance contributors are provided in the README.

Licensing & Compatibility

The provided README does not specify a software license. This lack of explicit licensing information may pose compatibility issues for commercial use or integration into closed-source projects.

Limitations & Caveats

The citation capture success rate of 76.44% indicates that data acquisition is not exhaustive. The research focuses on specific AI search platforms and prompt engineering techniques, and findings may not generalize to all search contexts. The absence of a declared license is a significant adoption blocker.