Institutional Data

Primary-Source Supply Chain and Electricity Demand Data

A point-in-time panel of weather-normalized US electricity demand and global logistics activity, built for systematic research.

Live production output

Request a sample extract

What we manufacture

01

EIA hourly demand

Public hourly metered electricity demand for eight US grid regions enters as the raw input.

02

Weather normalization

Monthly fitted coefficients remove heating and cooling load, leaving the demand attributable to underlying activity.

03

Sector split

A sector split isolates industrial demand from commercial and data center demand.

04

Scored series

Regional readings are baselined, scored, and combined into a GDP-weighted national series.

Public EIA hourly demand enters; a proprietary, patent-pending transformation produces the series: weather normalization with monthly fitted coefficients, then a sector split isolating industrial from commercial and data center demand. Patent application filed December 2025. Coefficient values and calibration parameters are not published.

The panel

185+ proprietary series across US electricity demand (industrial and commercial, 8 grid regions), a global disruption composite, 49 ports and chokepoints, trade corridors, country customs friction, and freight rate pressure. Delivered via REST API, CSV, and an MCP server listed in the official MCP Registry.

Series familySeriesCadence
US electricity demand, industrial and commercial (8 grid regions)32Hourly
Data center sub-region demand5Daily
Global disruption composite and pillars5Every 15 minutes
Ports and chokepoints49Every 2 hours
Airports59Every 15 to 30 minutes
Trade corridors10Every 2 hours
Country customs friction21Daily
Freight rate indices4Daily

Nowcasting

Official manufacturing series are monthly and publish roughly 4 to 5 weeks after the period they measure. SupplyMaven's corresponding series update hourly, so activity observed in the demand data today does not appear in an official print for 4 to 9 weeks.

Validation and integrity

Every registered signal clears a pre-registered statistical bar: Granger F-test on first-differenced, stationarity-checked series, p at or below 0.01, directional accuracy at or above 55 percent, reverse causality testing, and pandemic-supercycle exclusion. Null results are documented, not hidden.

Lead validated relationship

Commercial electricity demand leads total US manufacturing output by one month.

Directional accuracy77.8 percent
Granger F19.5
p-value0.00005
CausalityUnidirectional
RobustnessPermutation-tested

Full validation record available on request.

Source observations have been captured by SupplyMaven since April 2024. Composite index series are computed live in the current production platform from early 2026; earlier index history is methodology-consistent reconstruction computed from stored source observations and is labeled as such wherever delivered. SupplyMaven operates a no-retroactive-adjustment policy: published values are not restated.

Engagement path

Evaluation follows a structured path.

  1. 01Sample extract with data dictionary and full validation record
  2. 02Researcher evaluation on your timeline
  3. 0390-day paid pilot, credited in full against a first-year license
  4. 04Non-exclusive annual licensing with capacity-limited distribution

Licenses include quarterly methodology and domain-context sessions with the founder.

Scott Sage, Founder

scott.sage@supplymaven.com