Monitoring and Detecting Cultural Heritage Manipulation in Armenia

A Data-Driven Approach to Analyzing Russia’s Digital Influence on Cultural Narratives

Overview

This repository contains code, documentation, and resources for a research project investigating Russia’s manipulation and weaponisation of cultural heritage in Armenia. By combining data mining, Wikipedia analysis, semantic retrieval, and LLM-assisted annotation, this project aims to detect, trace, and analyze narrative shifts, coordinated editorial activity, and ideological framing around Armenian cultural heritage.

The goal is to build a transparent, reproducible analytical framework supporting academic research, cultural preservation, and evidence‑based policy-making.

Motivation

Cultural heritage is a tool of political power, especially in conflict regions. Manipulation on platforms like Wikipedia can influence:

• Public perception
• International discourse
• Historical narratives
• Cultural identity

This project provides computational tools to document such interventions.

Methodology

Data Acquisition

• Wikimedia API
• Web scraping of article HTML and talk pages
• Metadata extraction

Revision & Diff Analysis

• Successive revision comparison
• Framing and sentiment shifts
• Detection of ideological terminology

Semantic Article Retrieval

Using multilingual embeddings (e.g., Alibaba-NLP/gte-multilingual).

Narrative Analysis

LLM‑based annotations and topic modeling.

Repository Structure

.
├── README.md
├── data/
│   ├── raw/
│   └── processed/
├── notebooks/
│   ├── XXX.ipynb
│   └── YYY.ipynb
├── src/
│   ├── processing/
│   ├── analysis/
│   └── utils/
├── requirements.txt
└── LICENSE