ThermoStrife Documentation

Historical-uprisings climate companion to ThermoKourt. Tests whether positive temperature anomalies are over-represented on days of mass civil strife, using a case-crossover null built from matched non-event days at the same station and calendar month.

User guide

• Getting started
  • Installation
  • Verifying the installation
  • Your first run
• Pipeline guide
  • Stage 1 — Curated event panels
  • Stage 2 — Tiered weather-archive cascade
  • Stage 3 — Anomaly construction
  • Stage 4 — Case-crossover inference
  • Stage 5 — Reporting
• Statistical methods
  • 1 — Core question and primary test
  • 2 — Why a naive t-test is insufficient
  • 3 — Deconfounding the outdoor-opportunity confound
  • 4 — Pre-registered hypotheses
  • 5 — Multiple comparisons
  • 6 — Power
  • 7 — Limitations
  • 7b — Data-coverage limitations (intrinsic missingness)
  • 8 — References

Development

• Contributing
  • Development setup
  • Running the test suite
  • Linting
  • Building the docs locally
  • Commit conventions
  • Pull request workflow
• Continuous integration
  • tests.yml
  • docs.yml
  • release.yml
  • Branch protection
• Release process
  • 1. Pre-release checks
  • 2. Update version metadata
  • 3. Update the changelog
  • 4. Tag and push
  • 5. Post-release
  • Rolling back a broken release

API reference

• thermostrife.constants
• thermostrife.lookup
• thermostrife.inference
• thermostrife.viz

Project

• Changelog

Overview

ThermoStrife curates a 112-event panel of violent uprisings, revolutions, massacres, and prison revolts (1750–2026) and tests the heat-aggression hypothesis on the human-data side of the ThermoKourt programme. See the Statistical methods page for the pre-registered analysis plan and the Pipeline guide guide for end-to-end usage.

Quick links

• Source: github.com/zerotonin/thermostrife

• Issues: GitHub Issues

• Companion: ThermoKourt