Ask the Grid
Open live grid

Load

Actual and forecast electricity demand, from five-minute system totals to seven-day-ahead outlooks, by zone, area, and system, across every market in the catalog.

44 datasetspublished by ATG DataISONE13MISO7ERCOT6NYISO5PJM5CAISO4SPP4

Load datasets
DatasetRows
PJM7-Day Load ForecastATG Data·PJM·PJM's seven-day-ahead hourly load forecast by forecast area, refreshed through the day. Rows are keyed by target time (event_time) with the issue time in evaluated_at_utc; take the latest vintage per target hour via an argMax over evaluated_at_utc. Forecast areas match the very-short-term feed (utility names, RTO_COMBINED for the system total). This is the table for tomorrow's peak and the week-ahead demand outlook. Source: PJM Data Miner 2.5.0M2026
MISOActual Demand by RegionATG Data·MISO·Observed hourly MISO demand in MW for the Central, North, and South sub-regions. Region labels are normalized to uppercase; this is regional operational load, not a local-resource-zone series.216.4K2026
SPPActual System LoadATG Data·SPP·SPP's realized system-wide demand in MW at five-minute resolution, the actual-load counterpart to the short-term and mid-term load forecasts. Note that the feed republishes the same event_time on successive polls with a fresh ingested_at, so a read should collapse repeats on ingested_at before use.30.2K2026
CAISOCAISO Day-Ahead Load ForecastATG Data·CAISO·CAISO's day-ahead demand forecast, one row per forecast hour with a system total and one column per TAC area. 75,727 hours, 2018-01 through the next delivery day. Column-for-column identical to CAISO Load (caiso_load) on purpose: system_mw, pge_mw, sce_mw, sdge_mw, vea_mw, mwd_mw. Forecast-versus-actual is therefore a plain join on event_time with no reshaping on either side. CAISO publishes no issue-time column for this feed, so there is no vintage axis here — this is the day-ahead market run for each hour, not a curve of successive forecasts. CAISO's other horizons (two-day, seven-day, real-time) live in caiso.load_forecast_horizons and only begin in May 2026.75.9K2018
CAISOCAISO Hourly Actual LoadATG Data·CAISO·Actual hourly electricity demand in CAISO, one row per hour with a system total and one column per TAC area. 75,689 hours, 2018-01 to now. Columns: system_mw plus pge_mw, sce_mw, sdge_mw, vea_mw, mwd_mw. system_mw is CAISO's own CA ISO-TAC total. The five TAC areas are the CAISO balancing area's own partition. Wide rather than long, so "load right now" and "the 5 TAC area shares of it" are one row, not 6. Every market's load table in this catalog shares the system_mw column and the same hourly grain, so a cross-market demand comparison needs no reshaping.75.9K2018
ERCOTCurrent Supply, Demand & Intraday ForecastATG Data·ERCOT·ERCOT's public real-time supply and demand dashboard at five-minute resolution. This is the preferred dataset for a current actual-versus-forecast demand comparison: series='realized' reports observed demand_mw, series='intraday_forecast' reports demand_mw forecast for the same target event_time at successive source_time vintages, and series='forward_forecast' reports forecasted_demand_mw farther ahead. It also carries committed and available capacity. Operators, traders, and researchers use it to measure current forecast error, headroom, and how forecasts converge toward realized demand.6.3M2026
MISODay-Ahead Cleared DemandATG Data·MISO·Hourly day-ahead cleared MISO demand by sub-region, separated into fixed, price-sensitive, and virtual demand. Virtual demand is a financial market position and must not be reported as physical electricity consumption.163.1K2026
NYISODay-Ahead Zonal Bid LoadATG Data·NYISO·NYISO day-ahead zonal bid-load components by zone: energy bid load, bilateral load, price-cap load, virtual load, and virtual supply. Use it to compare bid-in demand with forecasts and realized load.2.6M2001
ISO-NEDemand DA HourlyATG Data·ISONE·ISO-NE's published Demand DA Hourly source feed. This catalog page exposes the physical schema, coverage, and latest source rows without inventing a normalized grain for publisher-shaped snapshot fields. Use the documented key and event time when comparing revisions or vintages.8.0K2026
ISO-NEDemand RT HourlyATG Data·ISONE·ISO-NE's published Demand RT Hourly source feed. This catalog page exposes the physical schema, coverage, and latest source rows without inventing a normalized grain for publisher-shaped snapshot fields. Use the documented key and event time when comparing revisions or vintages.2842026
ERCOTERCOT Hourly Actual LoadATG Data·ERCOT·Actual hourly electricity demand in ERCOT, one row per hour with a system total and one column per weather zone. 119,475 hours, 2013-01 to now. Columns: system_mw plus coast_mw, east_mw, far_west_mw, north_mw, north_central_mw, south_central_mw, southern_mw, west_mw. system_mw is ERCOT's own published total, and the eight zones sum to it to within 0.03 MW — so a zone share is simply zone_mw / system_mw. Wide rather than long, so "load right now" and "the 8 weather zone shares of it" are one row, not 9. Every market's load table in this catalog shares the system_mw column and the same hourly grain, so a cross-market demand comparison needs no reshaping.119.7K2013
NYISOFive-Minute Actual LoadATG Data·NYISO·NYISO preliminary real-time actual load by load zone at five-minute resolution. This is the higher-cadence load series for intraday grid movement, while hourly integrated actual load remains the settlement-quality hourly view.34.6M2001
ISO-NEFive-Minute System LoadATG Data·ISONE·ISO-NE's observed five-minute system load, native load, active demand response, and behind-the-meter-PV-adjusted totals. This is a market total and must not be copied into individual load zones.24.2K2026
ISO-NEHB DA Demand BidATG Data·ISONE·ISO-NE's published HB DA Demand Bid source feed. This catalog page exposes the physical schema, coverage, and latest source rows without inventing a normalized grain for publisher-shaped snapshot fields. Use the documented key and event time when comparing revisions or vintages.10.7M2024
ISO-NEHB DA Load Response OfferATG Data·ISONE·ISO-NE's published HB DA Load Response Offer source feed. This catalog page exposes the physical schema, coverage, and latest source rows without inventing a normalized grain for publisher-shaped snapshot fields. Use the documented key and event time when comparing revisions or vintages.01970
NYISOHourly Actual LoadATG Data·NYISO·NYISO integrated actual load by load zone at hourly resolution. The table covers the eleven NYISO zones and is the settlements-grade actual-load series used by the NYISO map cards; system actual load is computed by summing the zone rows.2.6M2001
ERCOTHourly Historical System LoadATG Data·ERCOT·Hourly weather-zone actual-load history for ERCOT, including the eight weather zones and a system total. Use it for zonal and historical hourly analysis; for the freshest current system-wide demand, prefer System-Wide Demand (ercot_system_wide_demand) after comparing latest timestamps. Analysts use this archive to study load shape and peaks, normalize demand against temperature, and contextualize price moves and reserve conditions. It comes from ERCOT report NP6-345-CD.119.8K2013
ISO-NEHourly System Load ForecastATG Data·ISONE·ISO-NE's hourly system load and net-load forecast with an explicit source_time vintage. Select the latest known source_time per target hour and do not sum forecast vintages.1.2K2026
PJMInstantaneous LoadATG Data·PJM·PJM metered load sampled every five minutes by area, where 'PJM RTO' is the whole-RTO system total and the remaining areas are the transmission-zone operating regions (AEP, COMED, DOM, PSEG, and so on). Because it posts within minutes, this is the live demand signal for PJM — system load, zonal load, and the shape of the daily peak — unlike the hourly metered-load tables that settle on a delay. Source: PJM Data Miner 2.383.9K2026
MISOLive Demand, Capacity, and Short-Term Reserve OutlookATG Data·MISO·MISO's current and forecast system demand, committed and available capacity, plus next-day short-term reserve requirements. Filter record_type before aggregating: supply_demand and next_day_short_term_reserve are distinct grains.14.8K2026
MISOLive System Load and ForecastATG Data·MISO·MISO's public live system-wide load bundle: five-minute actual load, hourly cleared load, and the medium-term hourly forecast. Filter series before use; the system total must not be copied into local-resource-zone values.942.7K2026
CAISOLoad ActualsATG Data·CAISO·Hourly actual electricity demand from CAISO OASIS, reported for the CAISO system total and TAC load areas such as PG&E, SCE, and SDG&E. This settlements-grade series is published after the operating interval rather than as a live five-minute reading. It supports forecast benchmarking, historical load shape, and aligned hourly net-load analysis. The official source is CAISO OASIS.2.0M2018
NYISOLoad ForecastATG Data·NYISO·NYISO hourly load forecast by load zone plus a NYISO system-total row, with source_time preserving forecast vintage. The application uses the latest vintage per target hour to compare expected load against realized hourly actuals.17.7M1999
ERCOTLoad Forecast (MTLF)ATG Data·ERCOT·ERCOT's mid-term load forecast (MTLF) for the Texas grid, giving the expected electricity demand for each of the next seven days at hourly resolution. The forecast is broken out by ERCOT's eight weather zones plus a system-wide total, and a new vintage is issued every hour, so you can also see how the outlook for a given day shifts as it approaches. Operators use this to plan generation and reserves, while traders and researchers use it to anticipate price and tightness conditions and to study how forecasts evolve and where they miss. The official source is ERCOT report NP3-565-CD, Seven-Day Load Forecast by Weather Zone.111.8M2017
CAISOLoad Forecast HorizonsATG Data·CAISO·CAISO load forecasts for the system and TAC load areas, including the near-term RTM 5-minute load forecast used by the application map. Each forecast row gives expected megawatt demand for a target interval and forecast label, so the data can be compared with actual load as conditions evolve. Operators and analysts use these horizons to understand expected ramping needs, assess forecast miss, and anticipate tight net-load periods before they show up in realized demand. The public source is CAISO Today's Outlook.38.5M2018
MISOLoad State Estimator by ZoneATG Data·MISO·Real-time demand in MW by MISO load-state-estimator zone. These operational zone identifiers are source-native and should not be assumed to match planning-zone or local-resource-zone geography without an explicit crosswalk.396.6K2026
ISO-NELoad Zone NCPCATG Data·ISONE·ISO-NE's published Load Zone NCPC source feed. This catalog page exposes the physical schema, coverage, and latest source rows without inventing a normalized grain for publisher-shaped snapshot fields. Use the documented key and event time when comparing revisions or vintages.1.7K2022
MISOMedium-Term Load ForecastATG Data·MISO·MISO's rolling ten-day hourly load forecast by region and local resource zone. The table is bitemporal: event_time is the delivery hour and source_time/init_date identify the forecast vintage, so select a vintage before evaluating forecast error.529.4K2026
SPPMid-Term Load Forecast (MTLF)ATG Data·SPP·SPP's mid-term load forecast in MW, running roughly seven days ahead, carried alongside the realized actual for the same interval so forecast error can be computed within the table. Bitemporal: each row records the delivery interval and the vintage the forecast was published at, retaining every successive forecast for a given hour.537.9K2026
ISO-NENetwork LoadATG Data·ISONE·ISO-NE's published Network Load source feed. This catalog page exposes the physical schema, coverage, and latest source rows without inventing a normalized grain for publisher-shaped snapshot fields. Use the documented key and event time when comparing revisions or vintages.4212020
NYISONYISO Hourly Actual LoadATG Data·NYISO·Actual hourly electricity demand in NYISO, one row per hour with a system total and one column per load zone. 101,978 hours, 2015-01 to now. Columns: system_mw plus west_mw, genesee_mw, central_mw, north_mw, mohawk_valley_mw, capital_mw, hudson_valley_mw, millwood_mw, dunwoodie_mw, nyc_mw, long_island_mw. system_mw is DERIVED as the sum of the eleven zones — NYISO's hourly integrated-load feed publishes no state total. It is therefore true by construction and is not independent corroboration of the zones. Wide rather than long, so "load right now" and "the 11 load zone shares of it" are one row, not 12. Every market's load table in this catalog shares the system_mw column and the same hourly grain, so a cross-market demand comparison needs no reshaping.220.8K2001
PJMPJM Hourly Actual LoadATG Data·PJM·Actual hourly electricity demand in PJM, one row per hour with a system total and one column per load area. 75,672 hours, 2017-12 to now. Columns: system_mw plus aep_mw, ap_mw, atsi_mw, dayton_mw, deok_mw, dominion_mw, duquesne_mw, ekpc_mw, mid_atlantic_mw, northern_illinois_mw. system_mw is DERIVED as the sum of the ten areas, but unlike NYISO it has been checked: against PJM's own published RTO total over 517 overlapping hours the mean difference is 0.009% and the maximum 0.938%. Wide rather than long, so "load right now" and "the 10 load area shares of it" are one row, not 11. Every market's load table in this catalog shares the system_mw column and the same hourly grain, so a cross-market demand comparison needs no reshaping.75.9K2017
PJMPJM Seven-Day Load ForecastATG Data·PJM·PJM's seven-day hourly demand forecast in long form: issued_at (when PJM published the forecast), event_time (the hour being forecast), area, and forecast_mw. 4.14M rows across 2,033 issue vintages and 25 areas, covering forecasts issued since 2026-05-21 and targets out to seven days ahead. Because issue time and target time are separate columns, this table can answer both 'what does PJM expect for tomorrow evening' and 'what did PJM expect for that hour twelve hours before it happened' — the second question needs the issue-time filter, and getting it wrong turns a forecast evaluation into hindsight. The area column mixes three levels of the same hierarchy: the RTO total (RTO_COMBINED), three regional aggregates (MID_ATLANTIC_REGION, SOUTHERN_REGION, WESTERN_REGION), transmission zones (AEP, COMED, DOMINION, ...) and Mid-Atlantic sub-zones (PSE&G/MIDATL, PECO/MIDATL, ...). Summing across areas double-counts, in some cases three times over.5.0M2026
MISOReal-Time Cleared Demand ForecastATG Data·MISO·MISO's system-wide hourly real-time cleared-demand forecast. It is a forecast series, not observed load; use demand_actual or live_load series='actual_five_minute' for realized demand.56.7K2026
ISO-NEReliability Region Load ForecastATG Data·ISONE·ISO-NE's published Reliability Region Load Forecast source feed. This catalog page exposes the physical schema, coverage, and latest source rows without inventing a normalized grain for publisher-shaped snapshot fields. Use the documented key and event time when comparing revisions or vintages.421.8K2022
ERCOTSCED Load Resource Data (60-Day)ATG Data·ERCOT·Per-resource, per-interval telemetry for ERCOT's controllable Load Resources — large loads that curtail or provide ancillary services — from the 60-Day SCED Disclosure Reports. Each row is one SCED run (roughly every five minutes) for one load resource and its QSE, showing its real power consumption alongside its maximum and low consumption, its high/low ancillary and dispatch limits (HASL/HDL, LASL/LDL), its SCED dispatch target (base point) and telemetered status, and how much of each ancillary service it was carrying (Reg-Up, Reg-Down, RRS, RRS-FFR, Non-Spin, ECRS). Because Load Resources are how demand-side capacity participates in ERCOT's real-time market and reserves, this dataset shows demand response actually happening at five-minute resolution — how much load was armed for responsive reserve, when it was deployed, and how it followed dispatch. Like the gen file it is published on a 60-day regulatory delay, making it a historical record for studying demand-side flexibility, reserve provision, and load participation in scarcity events. The official source is ERCOT report NP3-965-ER, 60-Day SCED Disclosure Reports.155.1M2021
ISO-NESeven Day ForecastATG Data·ISONE·ISO-NE's published Seven Day Forecast source feed. This catalog page exposes the physical schema, coverage, and latest source rows without inventing a normalized grain for publisher-shaped snapshot fields. Use the documented key and event time when comparing revisions or vintages.1.7K2022
SPPShort-Term Load Forecast (STLF)ATG Data·SPP·SPP's short-term load forecast in MW. Bitemporal: each row carries both the delivery interval (event_time) and the vintage the forecast was issued at (publish_time), so successive forecasts for the same interval are retained rather than overwritten. That makes the table usable for forecast-error and forecast-convergence analysis, not only for the latest view.119.6K2026
SPPSPP Load ForecastATG Data·SPP·SPP's demand forecast in long form: issued_at (when SPP published it), event_time (the hour or interval being forecast), horizon, and forecast_mw. 535,105 rows since 2026-05-20, out to seven days ahead. The horizon column separates SPP's two published products: 'short_term' (STLF, 24,599 vintages, five-minute targets out a few hours) and 'mid_term' (MTLF, 2,188 vintages, hourly targets out seven days). They are different forecasts at different resolutions and must never be mixed in one series. Issue time and target time are separate columns, so this table supports both 'what does SPP expect' and 'what did SPP expect N hours out'.639.9K2026
ISO-NESystem Load HourlyATG Data·ISONE·ISO-NE's published System Load Hourly source feed. This catalog page exposes the physical schema, coverage, and latest source rows without inventing a normalized grain for publisher-shaped snapshot fields. Use the documented key and event time when comparing revisions or vintages.40.8K2022
ERCOTSystem-Wide DemandATG Data·ERCOT·This dataset reports actual system-wide electricity demand across the ERCOT region of Texas, summed to a single total at 15-minute intervals. It comes from ERCOT's official Actual System Load by Weather Zone and system-wide reporting (report NP6-235-CD) and reflects settled demand rather than the faster, provisional real-time readings. Operators, traders, and researchers use it to study how Texas load rises and falls through the day and across seasons, to benchmark forecasts against what the grid actually consumed, and to anchor analyses of peak demand and stress events.15.5K2026
PJMVery Short-Term Load ForecastATG Data·PJM·PJM's very-short-term load forecast by forecast area, covering roughly the next two hours at fine granularity. Rows are keyed by their target time (event_time) with the issue time in evaluated_at_utc, so the latest vintage for each target comes from an argMax over evaluated_at_utc. Forecast areas use utility names (AE/MIDATL, DOMINION, and RTO_COMBINED for the system total). Use it for the immediate demand trajectory; for tomorrow and the week ahead use the 7-day forecast. Source: PJM Data Miner 2.15.5M2026
ISO-NEZonal Load 5mATG Data·ISONE·ISO-NE's published Zonal Load 5m source feed. This catalog page exposes the physical schema, coverage, and latest source rows without inventing a normalized grain for publisher-shaped snapshot fields. Use the documented key and event time when comparing revisions or vintages.28.8K2026
ISO-NEZonal Load Forecast 5mATG Data·ISONE·ISO-NE's published Zonal Load Forecast 5m source feed. This catalog page exposes the physical schema, coverage, and latest source rows without inventing a normalized grain for publisher-shaped snapshot fields. Use the documented key and event time when comparing revisions or vintages.28.4K2026