GrivavaLAB

Independent interdisciplinary research on correlations between space weather — solar activity, geomagnetic disturbances, cosmic rays — and human physiological, psychological, and behavioural patterns as reflected in large-scale open data. Rooted in the tradition of heliobiology and Chizhevsky's pioneering work, extended with modern data-science methods.

Latest Research — Completed

Epilepsy vs Space Weather: 4,076-Day Correlation Analysis

We analysed 11 years of daily Google Trends data for epilepsy against 7 space-weather variables (F10.7, Kp, Dst, solar wind, and more). The result: a weak but statistically significant positive correlation with F10.7 solar flux at lag +11 days (r = +0.1417, FDR-corrected). Not proof of a causal link — but interesting enough to dig deeper. Full report with interactive charts inside.

Read the full report →
Null Result — Completed

Laziness vs Space Weather: No Relationship Confirmed

Same pipeline, same 4,076 days — opposite outcome. An initial sunspot correlation (r = +0.102) passed FDR but failed all five robustness tests. The yearly r ≈ 0.71 turned out to be a textbook Simpson's Paradox. This report is why we do robustness batteries.

Read the full report →
Null Result

Cortisol Levels vs Space Weather

Full robustness analysis of cortisol levels searches across 4,076 days. No credible relationship with space weather detected.

Read the full report →
Null Result

Domestic Violence vs Space Weather

27 of 427 hypotheses FDR-significant — but all fail robustness (block bootstrap, subperiod stability, placebo). |r| < 0.061 for all variables. No credible relationship between domestic-violence Google searches and space weather.

Read the full report →

About the Project

Open Data & Reproducibility

GrivavaLAB combines publicly available space-weather archives and anonymised aggregate search-interest data (Google Trends) to test for statistically significant associations between solar/geomagnetic activity and population-level health indicators. All methods are documented and reproducible.

Scientific Tradition

Rooted in the tradition of Alexander Chizhevsky's pioneering work on solar cycles and mass behaviour, and contemporary studies linking geomagnetic disturbances with cardiovascular and neurological outcomes. We aim not to confirm a hypothesis, but to rigorously evaluate the evidence using modern data-science methods.

Open Science

All code, data pipelines, and intermediate results are documented and shared openly. Questions this fundamental deserve transparent, community-verifiable answers.

Interdisciplinary Scope

Bridging heliophysics, epidemiology, neuroscience, and data science to explore how space weather may influence human physiology and behaviour at population scale.

Data Sources

SourceDataCoverage
NASA OMNIWebSolar wind, IMF, geomagnetic indices1963–present, hourly
NOAA SWPCKp, Ap, Dst indices1932–present, 3-hourly
NASA DONKICME catalogues, flares, proton events2010–present
GOESX-ray flares, proton flux1998–present, 1-min
SILSOSunspot number1749–present
DSCOVR / ACEReal-time solar wind at L11997–present, 1-min
Google TrendsSearch interest for medical terms2004–present, daily

Methodology

Reproducible analysis pipeline:

Data ingestion Temporal alignment Correlation analysis FDR control

All analyses in Python (pandas, SciPy, statsmodels). Lagged correlations (−30 to +30 days). Benjamini–Hochberg FDR correction at α = 0.05. Seasonal decomposition (STL, 365-day) and surrogate-data testing separate genuine signals from periodic artefacts.

Current Status

First analysis completed — July 2026. We completed a cross-correlation study of epilepsy-related search interest vs 7 space-weather variables across 4,076 days (2015-03-18 → 2026-07-15). Google Trends 90-day windows were stitched into a continuous daily series. STL detrending, FDR correction, and 1,000-permutation surrogate testing were applied. Read the full report →

Updates & Research Log

View all articles →
  • 2026-07-18 resultsnull-result [Domestic Violence] Null result. Domestic Violence vs Space Weather — 27/427 FDR-significant initially, but all fail robustness. No credible relationship.
  • 2026-07-17 dataresultsnull-result [Cortisol Levels] Null result. Cortisol Levels vs Space Weather — No credible relationship found after full robustness testing.
  • 2026-07-17 resultsdata First correlation analysis completed. Epilepsy vs Space Weather: 4,076-day analysis — F10.7 solar flux shows weak positive correlation at lag +11 days (r = +0.1417, FDR-corrected).
  • 2026-07-17 robustnessnull-result Robustness analysis completed. Laziness vs Space Weather — initial sunspot signal (r = +0.102) failed all 5 robustness tests. Verdict: relationship not confirmed.
  • 2026-07-16 infra Server provisioned; GrivavaLAB launched.
  • 2026-07-10 data Google Trends pipeline for epilepsy-related terms deployed. Initial 11-year data pull completed.
  • 2026-06-28 pipeline OMNIWeb & SILSO ingestion scripts finalised. Temporal alignment tested on 2004–2024.
  • 2026-06-15 meta Project repository created. Core methodology outlined.
8
Data Sources
11+
Years Analysed
4
Pipeline Stages
4
Completed Studies