Trees, forests, and neural networks on governance data; dictionaries and large language models on Federal Reserve, IMF, and World Bank documents. Everything runs in claude.ai — no installation, no code.
Each opens in its own page. The datasets and documents are yours to download; the Day 1 and Day 2 pages put ready-to-run prompts next to a link into Claude.
Twenty-three country-year datasets built from the University of Gothenburg's QoG data — each with a one-click data and description download.
Open datasets → 32Central bank statements, IMF and World Bank reports, rating actions, and communiqués — the raw material for the Day 2 text-analysis labs.
Open documents → 1Prediction on governance data: decision trees, random forests, and neural networks. Full schedule plus every Day 1 slide deck to download.
Open Day 1 → 2Turning documents into data: dictionary sentiment and large language models on real Fed, IMF, and World Bank text. Schedule plus all Day 2 decks.
Open Day 2 → ∗The working space: copy a ready-made prompt, open Claude in a new tab, run it on your file, then interpret and verify the result.
Open the labs →