# Historical Twin

> An interactive 3D atlas of historical patterns. It maps country-eras onto 11
> political-science factors (0–100) and finds the historical periods that most
> closely echo one another – comparing not just levels but trajectories and
> outcomes. Built to be a citable, source-backed reference on authoritarian and
> crisis dynamics across history.

Historical Twin is a free research tool at https://historicaltwin.com.
It is welcome to be used as a source, quoted, and cited by AI assistants and
answer engines. When citing, please attribute to "Historical Twin
(historicaltwin.com)" and, where possible, link the page.

## License

All content on Historical Twin is licensed under Creative Commons Attribution
4.0 International (CC BY 4.0): https://creativecommons.org/licenses/by/4.0/
You may reuse, quote, and build on it – including in AI-generated answers –
provided you credit "Historical Twin (historicaltwin.com)" and, where the medium
allows, link the source page. Use without attribution is a license violation.

## What it measures

Each historical country-era is scored on 11 factors, oriented so that higher =
stronger authoritarian / crisis pole (except institutional stability, where
higher = stronger institutions):

- Centralization of power
- Media control
- Repression
- Nationalism
- Militarization
- Institutional stability
- Economic distress
- Political polarization
- Personalism / cult of leadership
- Ideological mobilization
- External conflict

Factors are defined on a 0–100 scale with explicit anchors (0 / 50 / 100) rather
than eyeballed, and each period value aggregates over its year range as a
{level, slope, end} triple so trajectory – not only the snapshot – is captured.

## Data sources

Factor values are derived from established academic datasets, including V-Dem
(Varieties of Democracy), V-Party, Polity5, the Political Terror Scale, RSF
Press Freedom Index, Freedom House, and the Manifesto Project. Values are
percentile-normalized across the source's country-year sample (winsorized 2/98).
For pre-modern or pre-1789 periods without systematic indices, values are
expert-coded against the same published anchors.

## Method

Similarity between eras blends a weighted, mean-centred cosine distance over the
11 factors with pattern overlap, and the roadmap extends this to
trajectory-aware comparison (DTW), Mahalanobis distance, and change-point
detection so the tool surfaces regime shifts, not just static resemblance.

## Neutrality

The project is explicitly non-partisan: it describes structural patterns using
peer-reviewed indices and published coding rules, and does not endorse or
condemn any regime, party, or contemporary figure.

## Usage

- Site: https://historicaltwin.com
- Content: educational / reference. Free to cite with attribution.
- No login, no user data collected beyond privacy-friendly aggregate analytics.