Data Center Intelligence sample layers

Schema and provenance

Generated synthetically (seeded, reproducible) for the decision-support prototype. Point layers (data centers, cloud regions, hyperscalers, power plants, internet exchanges, tax incentives, latency) carry coordinates and layer-specific attributes; line layers (transmission, fiber) carry routes; choropleth layers (wildfire, drought, flood, earthquake, water, land, population, workforce) carry an illustrative per-metro score. No feature describes a real facility, hazard value, or incentive. Per-layer GeoJSON lives under data/layers/{date}/{layer_id}.geojson and mirrors the engine's immutable snapshots; the time series is append-by-design and excluded from checksum verification.

The published CSV is the accumulating time series (timeseries.csv): one row per snapshot date with layer_count and total_features. The substantive data is the per-layer GeoJSON, one file per layer per date, downloadable from the dashboard as you toggle layers on.

Layer groups and fields:

Infrastructure — data_centers (mw, operator, metro), cloud_regions (provider, azs), hyperscalers (company, planned_mw), power_plants (fuel, mw), transmission (kv, from, to), fiber (carrier, route), internet_exchanges (participants, capacity_tbps).

Climate & hazard — wildfire (wildfire_risk), drought (drought_index), flood (flood_risk), earthquake (pga_g), water (water_stress).

Human factors — land (developable_pct), population (population), workforce (tech_workforce_k), tax_incentives (program, value_pct), latency (p50_ms_to_nearest_hub).

Immutable artifacts are the per-day snapshots and their metadata.json manifest (with SHA-256 checksums). The time series grows daily by design and is therefore skipped by verify:datasets.