Substation Scout

How Substation Scout gets its data

Substation Scout republishes public data. It does not survey sites, run power-flow studies or hold any non-public utility information. This page lists every source, how often it is read, what is done to it, and where it stops. Every figure on the site carries its source and retrieval date, so any number can be traced back to a file listed here.

Wording rules

Every page follows the same rules, and automated tests check the published text against them:

  • State what a source contains, never what it proves.
  • A withdrawal is reported as a request leaving the queue, and nothing more is read into it.
  • Indirect signals are indicators, not determinations.
  • Every figure carries its source and retrieval date, or it is not published.
  • A field the source leaves blank or ambiguous is shown as "Not reported", not filled in.

Substation Scout is a screening aid. It is not a system impact study, and nothing here replaces a study by the relevant operator or utility.

Interconnection queues (Queue Wire and the region pages)

Three regions are covered: PJM, ERCOT and BPA. Each is read from the file the operator itself publishes:

  • PJM: the serial-queue XML (PlanningQueues.xml) and the cycle-process export, both from pjm.com. The two views overlap, so rows that appear in both are merged into one request.
  • ERCOT: the newest monthly GIS Report workbook on ercot.com. ERCOT publishes monthly, so its figures can trail the other regions by about a month.
  • BPA: the interconnection queue workbook on bpa.gov. Generation rows go to Queue Wire; load rows go to the Large Load Registry.

The pipeline runs weekly, on Monday mornings US Eastern time. Each run is compared with the previous one, and a change is logged when a request appears, changes status, changes capacity, or reaches a terminal status (withdrawn, cancelled, terminated and similar wording, matched regardless of case). Running the same file twice logs nothing.

A region's first run is a baseline, not a week of activity: its rows are first sightings and are never published as new requests.

If a live PJM or ERCOT download fails, the pipeline falls back to the last good file, marks the run degraded, and does not publish. The site keeps the previous week's data rather than showing old numbers as new.

A withdrawal records that a request left the queue. It says nothing about injection capacity at that location.

Large Load Registry

The registry lists the large-load (LL) rows from the same BPA workbook: requests to connect load, not generation, to the Federal Columbia River Transmission System. Requested megawatts are reproduced as filed.

BPA publishes no end-use field, so no request is classified as a data center. The project name is reproduced as filed and the reader can judge it.

About one in six load rows has no state. Several counties in the BPA area (Benton, Columbia, Grant, Douglas, Lincoln) exist in both Washington and Oregon, so the state is never inferred from the county. Those rows read "Not reported".

Substation Expansion Tracker

The tracker reads the meeting materials of PJM's Transmission Expansion Advisory Committee (TEAC) and its Sub-Regional RTEP committees. Each PDF is stored once, with its URL, meeting date and retrieval time.

Substation names, costs, in-service dates and statuses are extracted from each PDF by a language model, which is instructed to copy values exactly as printed. Every extracted name is then checked against the PDF's own text layer. A name that is not printed in the text, for example one read only from a map or one-line diagram, is withheld rather than published. The extraction is not reviewed by a person.

Tracker names are not matched to substation location datasets. A tracker page links to a queue substation page only when a single candidate matches on the transmission owner or state.

Site Scanner

Site Scanner screens one address against public layers. Each layer, with the vintage it carries, is listed below this text. Notes on how they are used:

  • Addresses are geocoded with the US Census geocoder, which also supplies the county and place codes used for joins.
  • Substations come from HIFLD, at 69 kV and above. The original HIFLD portal closed in 2025; the layer is read from a public mirror and its vintage is the newest source date in the data.
  • Queue data in a report is joined to a substation only on an exact normalized name with the same state and county. There is no fuzzy matching: a wrong substation's queue data is worse than none.
  • Local moratoria come from the CC-BY-4.0 Moratorium Data 2026 dataset, pinned to a fixed commit so a report can be reproduced, and joined by Census code, never by place name. A listed instrument is not a determination that a site is covered.
  • Flood and watershed lookups are point-in-polygon checks against FEMA and USGS layers. They are not a determination by FEMA, the USGS or a utility.

Reports are deleted from the server within 24 hours of being generated.

Natural gas (Site Scanner)

Many large projects will need on-site generation until the grid catches up, so each report also screens the gas network near the site. It reports three things, each as what the source contains:

  • Gas distribution company: the gas utility territory polygon the address falls in. When no polygon contains the point, the report says so; that is a statement about the dataset, not a finding that the area has no gas utility.
  • Transmission pipelines within 25 miles: from the EIA natural gas interstate and intrastate pipeline layer, read from the copy Esri republishes in its Living Atlas. That copy is newer than the HIFLD copy of the same EIA data, which stopped updating in 2021, and it carries a status field.
  • Pipeline facilities within 15 miles: compressor stations, receipt and delivery points, storage fields and processing plants, from HIFLD.

The pipeline layer is compiled by EIA from federal and state sources at varying scale. It is not survey-grade: a line on the map is not a locate, and it carries no diameter, pressure or capacity. The facility points and the distribution territories exist only as HIFLD Open snapshots; the HIFLD portal closed in 2025, and the most complete public copy was last edited in February 2021, with older record dates. A pipeline near a site is not gas service to it.

Site Scanner layers and vintages

LayerSourceVintage
Substations (69 kV and above)HIFLD, public mirror2022‑06 vintage, checked 2026-08-31
Electric retail service territoriesHIFLD / EIA2023‑08 vintage, checked 2026-08-31
Gas transmission pipelinesEIA, via Esri Living AtlasLayer last edited 2025‑12, checked 2026-10-16
Gas facilities and gas utility territoriesHIFLD Open mirror2021 edition, checked 2026-10-16
Watersheds (HUC-8)USGS Watershed Boundary DatasetQueried live per report
Flood hazard zonesFEMA National Flood Hazard LayerQueried live per report
Geocoding and Census geographiesUS Census Bureau geocoderQueried live per report
Local moratoriaMoratorium Data 2026 (CC-BY-4.0)Pinned commit, stated in each report