Advanced Business Intelligence with OSINT and AI · goldenowl.ai
This repository is Golden Owl's open home for the datasets behind our research. Whenever we run an investigation and collect or generate data, we publish the underlying dataset here so analysts, journalists, researchers, and partners can access it, verify our findings, and build on them.
The goal is simple: one repository for all of our research data. Each study lives in its own folder; we do not create a separate repository per project, and this central README is the single place that documents what each folder contains.
golden-owl-research-data/
├── README.md ← you are here (central index)
└── <research-slug>/ ← one folder per research project
└── <data files> ← e.g. results.json
| Research | Folder | Description |
|---|---|---|
| Russian Influence Architecture (War Against Ukraine) | osint-russia-ukraine-influence-2026 | Multilingual OSINT dataset of 2026 web/news items analyzed for pro-Kremlin narratives across 9 languages. |
More datasets will be added here as new research is published.
- Each research folder contains the raw data file(s); this central README is the index describing them.
- Data files are typically JSON. Large files can be downloaded directly or accessed via the raw file URL.
- Collection scope and field definitions vary by study — see the dataset notes in the table above before analysis.
Golden Owl combines OSINT methodologies, network science, narrative analysis, computational linguistics (NLP), geopolitical intelligence, human verification, and strategic risk assessment. Datasets are produced through automated, large-scale, multilingual collection and AI-assisted analysis, always with analyst review. Specific tooling and providers are intentionally omitted; the focus is on transparent, reproducible outputs.
Unless stated otherwise in a dataset's own README, these datasets are shared for research, analysis, and educational purposes. They aggregate publicly available online content together with automated analytic labels; analytic labels reflect AI-assisted assessment and should be treated as analytical signal, not adjudicated fact. Inclusion of a source does not imply any claim about that source beyond what the dataset's methodology describes.
Golden Owl — goldenowl.ai