Data

Numerical output from each logbook entry, stored as compressed NumPy archives (.npz).

Naming convention

{entry_number}_{short_name}.npz — one file per logbook entry.

All data files use np.random.default_rng(2026) unless stated otherwise in the corresponding logbook entry.

File	Entry	Description	Size
`001_aipp_convergence.npz`	001	AIPP per N per realisation (14 sample sizes × 200 realisations); 95th-percentile thresholds for 5 noise models	25 KB
`002_sigma_sensitivity.npz`	002	AIPP distributions under 4 perturbation conditions at N = 50 and N = 200 (300 realisations each)	20 KB
`003_powerlaw_thresholds.npz`	003	AIPP distributions for all 10 null models (300 realisations); finite-N bias fit coefficients (a, b, AIPP_inf)	26 KB
`004_delta_min.npz`	004	Null distributions of variance slope and autocorrelation for all 10 models (300 realisations × T = 200); calibrated delta_min values; sanity-check signal arrays	7.9 MB
`005_comparison_fom.npz`	005	Controlled comparison: 20 clocks × 200 steps; per-clock chi2, Huber, and IC arrays	121 KB

Entry 004 is the largest file (~8 MB) because it stores the full null distributions across all ten noise models. This is within the GitHub size limit but approaches the threshold where git-lfs or Zenodo hosting should be considered.
WP2 simulation data (when generated) will likely exceed 5 MB per scenario. At that point, consider migrating to git-lfs or archiving on Zenodo with DOI.
To load: data = np.load('data/001_aipp_convergence.npz'); data['aipp_N50'] etc.
Metadata fields (seed, n_realisations, etc.) are stored as scalar arrays within each .npz file.

All data can be regenerated from the scripts in scripts/: