Methodology

Every node in the atlas is a dated record: a concentration figure, a location, a written fault line, named actors, sources. No figure is fabricated. When a value is approximate it carries the ≈ sign; when sources diverge it is flagged. Records favour institutional sources (EIA, USGS, IEA, World Nuclear Association, DOE, SEC filings) and dated specialist press.

The IF sums up a stratum's structure in a score from 0 to 100. It combines three readings of the graph, not market volumes, but the shape of the dependency:

IF = 100 × (0.5·C + 0.3·S + 0.2·G). The index is comparative: it serves to read strata against one another and to track their evolution as records accumulate, not to predict a rupture. It will move with the graph, by design: the more the observatory documents, the more honest the index becomes.

Scenarios propagate a hypothetical disruption along the graph's veins. The model is deliberately simple and transparent: it follows dependencies downstream only (supply disruption, not loss of customers); a node's impact is the weighted share of its broken supplies, attenuated by a factor of 0.8 for each “redundancy” supplier still standing, capped at two, nominal redundancy never absorbs everything; the tiers (T+7d → T+1yr) express propagation distance, not a calendar. It is not a forecast: it is a reading of the network's shape, where a disruption can reach, and by which paths.

The seismograph catalogues real shocks, geopolitical, climatic, industrial, and projects them onto the structure: which nodes were hit, along which paths the disruption spread. This is the product's thesis: a system reads better under stress than at rest.