HUMANITYVECTOR

Methods

Derived metric laboratory

Every derived metric is a declarative formula: which observed series it reads, how each is transformed, how they are combined, what happens when a component is missing, and how the result is rounded. The spec is versioned; a change is a new version with a new effective date, and values computed under the old version stay in the history. Each stored value records the exact component observation ids it was computed from, and a weekly check recomputes every value from those ids.

Derived

21 specs registered, 19 with computed values.

Derived metricVersionComponentsFormula in plain wordsCoverageLast computed
Constraint centrality: aluminumconstraint-centrality-aluminum1.0none (assessment)The custom function constraint_centrality over the whole component dataset, for each geography and period. A period with any component missing produces no value. Rounded to 3 decimals.1 values, 1 geography, 2026-09-07 to 2026-09-072026-09-07 22:04 UTC
Constraint centrality: cementconstraint-centrality-cement1.0none (assessment)The custom function constraint_centrality over the whole component dataset, for each geography and period. A period with any component missing produces no value. Rounded to 3 decimals.1 values, 1 geography, 2026-09-07 to 2026-09-072026-09-07 22:04 UTC
Constraint centrality: copperconstraint-centrality-copper1.0none (assessment)The custom function constraint_centrality over the whole component dataset, for each geography and period. A period with any component missing produces no value. Rounded to 3 decimals.1 values, 1 geography, 2026-09-07 to 2026-09-072026-09-07 22:04 UTC
Constraint centrality: steelconstraint-centrality-electrical-steel1.0none (assessment)The custom function constraint_centrality over the whole component dataset, for each geography and period. A period with any component missing produces no value. Rounded to 3 decimals.1 values, 1 geography, 2026-09-07 to 2026-09-072026-09-07 22:04 UTC
Constraint centrality: electricityconstraint-centrality-electricity-system1.0none (assessment)The custom function constraint_centrality over the whole component dataset, for each geography and period. A period with any component missing produces no value. Rounded to 3 decimals.1 values, 1 geography, 2026-09-07 to 2026-09-072026-09-07 22:04 UTC
Constraint centrality: transformersconstraint-centrality-power-transformer1.0none (assessment)The custom function constraint_centrality over the whole component dataset, for each geography and period. A period with any component missing produces no value. Rounded to 3 decimals.1 values, 1 geography, 2026-09-07 to 2026-09-072026-09-07 22:04 UTC
Constraint centrality: semiconductorsconstraint-centrality-semiconductor-fab1.0none (assessment)The custom function constraint_centrality over the whole component dataset, for each geography and period. A period with any component missing produces no value. Rounded to 3 decimals.1 values, 1 geography, 2026-09-07 to 2026-09-072026-09-07 22:04 UTC
Copper production top-3 country concentrationcopper-top3-concentration1.0The custom function top3_share_of_world over the whole component dataset, for each geography and period. A period with any component missing produces no value. Rounded to 1 decimal.2 values, 1 geography, 2024 to 20252026-09-07 22:04 UTC
Corporate physical economy pulsecorporate-physical-economy-pulse1.0none (assessment)The custom function corporate_pulse over the whole component dataset, for each geography and period. A period with any component missing produces no value. Rounded to 2 decimals.1 values, 1 geography, 2026-08-02 to 2026-08-022026-09-07 22:04 UTC
Corporate pulse: ai infrastructure capex growthcorporate-pulse-ai-infrastructure1.0none (assessment)The custom function corporate_pulse over the whole component dataset, for each geography and period. A period with any component missing produces no value. Rounded to 2 decimals.1 values, 1 geography, 2026-07-03 to 2026-07-032026-09-07 22:04 UTC
Corporate pulse: global logistics capex growthcorporate-pulse-global-logistics1.0none (assessment)The custom function corporate_pulse over the whole component dataset, for each geography and period. A period with any component missing produces no value. Rounded to 2 decimals.no values computed yetnever
Corporate pulse: grid investment capex growthcorporate-pulse-grid-investment1.0none (assessment)The custom function corporate_pulse over the whole component dataset, for each geography and period. A period with any component missing produces no value. Rounded to 2 decimals.1 values, 1 geography, 2025-12-31 to 2025-12-312026-09-07 22:04 UTC
Corporate pulse: heavy industry capex growthcorporate-pulse-heavy-industry1.0none (assessment)The custom function corporate_pulse over the whole component dataset, for each geography and period. A period with any component missing produces no value. Rounded to 2 decimals.no values computed yetnever
Corporate pulse: semiconductor capacity capex growthcorporate-pulse-semiconductor-capacity1.0none (assessment)The custom function corporate_pulse over the whole component dataset, for each geography and period. A period with any component missing produces no value. Rounded to 2 decimals.1 values, 1 geography, 2026-07-03 to 2026-07-032026-09-07 22:04 UTC
Independently survivable human worldsindependently-survivable-human-worlds1.0none (assessment)The custom function dated_assessment over the whole component dataset, for each geography and period. A period with any component missing produces no value.1 values, 1 geography, 2026 to 20262026-09-07 22:04 UTC
Labor hours per CPI basket, indexed to 2007labor-hours-cpi-basket-indexed1.0The custom function labor_hours_cpi_basket over the whole component dataset, for each geography and period. A period with any component missing produces no value. Rounded to 4 decimals.244 values, 1 geography, 2006-03-01 to 2026-07-012026-09-07 22:04 UTC
Labor hours per 100 gallons of gasolinelabor-hours-per-100-gallons-gasoline1.0The custom function labor_hours_gasoline_monthly over the whole component dataset, for each geography and period. A period with any component missing produces no value. Rounded to 3 decimals.246 values, 1 geography, 2006-03-01 to 2026-08-012026-09-07 22:04 UTC
Labor hours per MWh of residential electricitylabor-hours-per-mwh-residential1.0Electricity_price_usd_per_mwh for USA multiplied by 10 divided by hourly_earnings for USA, for each geography and period. A period with any component missing produces no value. Rounded to 3 decimals.243 values, 1 geography, 2006-03-01 to 2026-05-012026-09-07 22:04 UTC
Lithium production top-3 country concentrationlithium-top3-concentration1.0The custom function top3_share_of_world over the whole component dataset, for each geography and period. A period with any component missing produces no value. Rounded to 1 decimal.2 values, 1 geography, 2024 to 20252026-09-07 22:04 UTC
Patent applicationspatent-applications1.0The sum of IP.PAT.RESD and IP.PAT.NRES, for each geography and period. A period with any component missing produces no value.4214 values, 167 geographies, 1980 to 20212026-09-07 22:04 UTC
Rare earth production top-3 country concentrationrare-earth-top3-concentration1.0The custom function top3_share_of_world over the whole component dataset, for each geography and period. A period with any component missing produces no value. Rounded to 1 decimal.2 values, 1 geography, 2024 to 20252026-09-07 22:04 UTC

What a spec records

A spec names its components by metric slug, each with an optional geography pin, lag, transform (level, percent change, natural logarithm, index to a base period, per capita), scale factor and positivity constraint. It names the combination (sum, ratio, product, difference, weighted sum, or a named function kept in code), an optional trailing window, an optional normalization (z-score or percentile against the geography’s own history), the missing-data policy (skip the period, or partial with a minimum component count and the coverage recorded), the unit, the default geography and the rounding.

Every computed value is stored with the spec version, a hash of the spec, the formula sentence, the list of component observation ids and values it used, the component release labels and the coverage. The reproducibility check reads those ids back, re-evaluates the spec over exactly those rows and compares. A mismatch marks the methodology source unhealthy.

Changing anything that affects the numbers is a new version. The engine refuses to write under a version whose stored hash differs from the code, so a silent edit cannot happen; a bumped version gets its own method version row with the date it took effect, and every value is written again under it while the earlier rows stay in the revision history.

No composite yet. Nothing here weights dimensions against each other. A spec that would combine unlike units is not published until its components and weights are written down and defended on its own page. See methodology versioning on the methods page.