Market Aggregates

What are Market Aggregates for?

Market aggregates, especially used with weighting such as capacity weighting, are useful proxies to estimate how differences in forecasts - either between models or between different forecasts of the same model - impact energy production & consumption.

Imagine you are comparing the latest 12PM forecast of EPT against the 6 AM forecast. In Deckenpfronn, Germany, there is a huge difference in prediction for 100m wind speed. But since there is no wind park in Deckenpfronn this difference in wind speed will have no impact at the energy production at all. However, differences in wind speed at for example Reussenkoge Wind Farm, with a production capacity of over 250 MW, will have significant impacts on the energy output.

This is why we offer to compute capacity weighted averages for both solar and wind parameters. The averages are computed as

$V_{avg, weighted} =\sum_i \dfrac{C_i}{C_{tot}} \cdot V_i$

Where $V_{i}$ is the value of the variable (e.g. wind speed) at a specific location (e.g. a wind farm), $C_i$ is the production capacity at that location and $C_{tot} = \sum_i C_i$ is the total production capacity of the selected region.

Output MW Mode

By adding unit=mw to the GET /v1/forecast/market-aggregate endpoint, the query engine applies generic power curves to weather forecasts and returns predicted megawatts (MW) instead of raw weather values. This is useful for comparing different models and quick MW estimates across many European zones.

Looking for actual production forecasts? MW output here provides a model-specific forecast of potential production via power curves. For a forecast of actual production, use the Power Forecast — an end-to-end model trained on real generation data, significantly more accurate. Currently available for Germany, France, United Kingdom, Belgium, and Netherlands (Solar, Wind).

MW mode is available on both endpoints:

• GET /v1/forecast/market-aggregate?unit=mw&...

• POST /v1/forecast/data with weighting.unit=mw in the body

MW mode:

• Activated by setting unit=mw (GET query parameter) or weighting.unit="mw" (POST body)

• Three weighting types are supported. In MW mode the variables field does not change the result — each weighting always produces the same MW columns:

  • weighting=wind_capacity → returns wind_onshore_mw, wind_offshore_mw (or wind_total_mw for combined-only zones such as GB)

  • weighting=solar_capacity → returns solar_mw

  • weighting=population → returns load_mw (electricity-demand forecast)

• To retrieve weather values at a specific height (e.g. 10m wind), use weather mode instead (omit unit=mw)

• Restricts geo to market_zone or country_key (no polygon or bounding box)

• Only available for zones listed under MW-Capable Zones below

• Weighting is auto-derived from the requested variable on the GET endpoint (wind speed → wind_capacity, shortwave flux → solar_capacity, 2m temperature → population)

• Wind and solar can be selected simultaneously on the GET endpoint; the backend issues separate queries per variable and merges

• When merge_zones_or_countries is enabled, MW values are summed across zones

Optional MW debiasing

MW output is raw by default if you omit debias. Examples on this page pass debias=true so new integrations get the walk-forward correction. Set debias=true on GET /v1/forecast/market-aggregate to apply the same walk-forward additive MW bias correction used by Power Forecast: eight weeks of history for wind and four weeks for solar and load to estimate the bias for each init time and lead; that bias is subtracted from the current forecast, then the window moves ahead.

GET /v1/forecast/market-aggregate?unit=mw&debias=true&...

debias=true requires unit=mw and market_zones. It is not supported with country_keys or weather units.

For POST /v1/forecast/data, use the corresponding request-body field:

{
  "weighting": {"type": "wind_capacity", "unit": "mw"},
  "mw_walkforward_debias": true
}

Historical periods with negative day-ahead prices are left out of the bias estimate. Omitting the option—or setting it to false—returns the unchanged raw MW forecast.

Example: Wind MW Forecast

Curl

curl -X GET "https://query.jua.ai/v1/forecast/market-aggregate\
?models=ept2\
&init_time=2025-10-22T00:00:00Z\
&weighting=wind_capacity\
&market_zones=DE\
&variables=wind_speed_at_height_level_100m\
&max_prediction_timedelta=72\
&unit=mw\
&debias=true\
&include_time=true\
&format=json" \
  -H "X-API-Key: YOUR_API_KEY" \
  -H "Accept: application/json"

Python

import requests

url = "https://query.jua.ai/v1/forecast/market-aggregate"
headers = {
    "X-API-Key": "YOUR_API_KEY",
    "Accept": "application/json"
}
params = {
    "models": "ept2",
    "init_time": "2025-10-22T00:00:00Z",
    "weighting": "wind_capacity",
    "market_zones": ["DE"],
    "variables": ["wind_speed_at_height_level_100m"],
    "max_prediction_timedelta": 72,
    "unit": "mw",
    "debias": True,
    "include_time": True,
    "format": "json",
}

response = requests.get(url, headers=headers, params=params)
data = response.json()

Response

{
    "model": ["ept2", "ept2", ...],
    "init_time": ["2025-10-22T00:00:00+00:00", ...],
    "prediction_timedelta": [0.0, 1.0, ...],
    "time": ["2025-10-22T00:00:00+00:00", ...],
    "market_zone": ["DE", "DE", ...],
    "wind_onshore_mw": [20109.78, 18485.41, ...],
    "wind_offshore_mw": [4933.67, 4880.13, ...]
}

Example: Load (Demand) Forecast

A population-weighted electricity-demand forecast in MW, produced by applying a temperature-driven load curve to 2m air temperature.

Curl

curl -X GET "https://query.jua.ai/v1/forecast/market-aggregate?models=ept2&init_time=latest&weighting=population&market_zones=DE&max_prediction_timedelta=72&unit=mw&debias=true&include_time=true&format=json" \
  -H "X-API-Key: YOUR_API_KEY" \
  -H "Accept: application/json"

Python

import requests

url = "https://query.jua.ai/v1/forecast/market-aggregate"
headers = {
    "X-API-Key": "YOUR_API_KEY",
    "Accept": "application/json",
}
params = {
    "models": "ept2",
    "init_time": "latest",
    "weighting": "population",
    "market_zones": ["DE"],
    "max_prediction_timedelta": 72,
    "unit": "mw",
    "debias": True,
    "include_time": True,
    "format": "json",
}

response = requests.get(url, headers=headers, params=params)
data = response.json()

Response

{
    "model": ["ept2", "ept2", ...],
    "init_time": ["2026-05-28T06:00:00+00:00", ...],
    "prediction_timedelta": [0.0, 1.0, ...],
    "time": ["2026-05-28T06:00:00+00:00", ...],
    "market_zone": ["DE", "DE", ...],
    "load_mw": [58525.5, 60100.2, ...]
}

The same query expressed against POST /v1/forecast/data:

payload = {
    "models": ["ept2"],
    "geo": {"type": "market_zone", "value": ["DE"]},
    "init_time": "latest",
    "variables": ["air_temperature_at_height_level_2m"],
    "prediction_timedelta": {"start": 0, "end": 72},
    "weighting": {"type": "population", "unit": "mw"},
    "mw_walkforward_debias": True,
    "group_by": ["model", "init_time", "prediction_timedelta", "time"],
    "aggregation": ["avg"],
    "include_time": True,
}

Load forecasts are available for ~48 zones today. Query GET /v1/forecast/market-aggregate/mw-zones and read the load key for the current list.

Example: Get Available MW Zones

curl -X GET "https://query.jua.ai/v1/forecast/market-aggregate/mw-zones" \
  -H "X-API-Key: YOUR_API_KEY"

{
    "wind": ["AT", "BA", "BE", "BG", "CH", "CY", "CZ", "DE", ...],
    "wind_combined": ["GB"],
    "wind_onshore_only": ["AT", "BA", "BG", "CH", "CY", "CZ", ...],
    "solar": ["AT", "BA", "BE", "BG", "CH", "CY", "CZ", "DE", ...],
    "load": ["AL", "AT", "BA", "BE", "BG", "CH", "CY", "CZ", "DE", ...]
}

The keys break down as:

• wind — zones that support wind_capacity weighting; output may include wind_onshore_mw and/or wind_offshore_mw.

• wind_combined — zones (today: GB only) that store wind capacity as a single combined value rather than split into onshore/offshore; output is wind_total_mw.

• wind_onshore_only — subset of wind zones that only have onshore capacity data fitted; output is wind_onshore_mw only.

• solar — zones that support solar_capacity weighting; output is solar_mw.

• load — zones that have a fitted population-weighted load curve; output is load_mw when weighting=population and unit=mw.

MW-Capable Zones

The list is dynamic and grows over time as new zones are onboarded. Always query the mw-zones endpoint for the current set rather than hard-coding a list.

MW Column Names

When MW mode is enabled, output columns are renamed:

Raw Column Name	Display Name	Unit	Weighting
wind_onshore_mw	Wind Onshore MW	MW	wind_capacity
wind_offshore_mw	Wind Offshore MW	MW	wind_capacity
wind_total_mw	Wind Total MW	MW	wind_capacity (combined-only zones)
solar_mw	Solar MW	MW	solar_capacity
load_mw	Load MW	MW	population

Variable-to-Weighting Mapping

When not using MW mode, the weighting field is applied based on the variable:

Variable	Weighting
wind_speed_at_height_level_100m	wind_capacity
wind_speed_at_height_level_10m	wind_capacity
surface_downwelling_shortwave_flux_sum_1h	solar_capacity
air_temperature_at_height_level_2m	population

Aggregation Periods and Column Naming

Daily and weekly aggregation is requested as a group_by entry on POST /v1/forecast/data — for example time__to_start_of(day) or time__to_start_of(week). The GET /v1/forecast/market-aggregate endpoint has no aggregation-period parameter (temporal_resolution resamples; it is not the same as daily/weekly bucketing).

Market aggregate responses use an avg__ prefix on variable columns. The response time column stays named time in every case — only the row count and the snapped timestamps change:

• Hourly (group_by includes time): avg__wind_speed_at_height_level_100m, time column time

• Daily (group_by includes time__to_start_of(day)): avg__wind_speed_at_height_level_100m, time column still time (timestamps snapped to day start)

• Weekly (group_by includes time__to_start_of(week)): avg__wind_speed_at_height_level_100m, time column still time (timestamps snapped to week start)

In MW mode, columns use entirely different names (see MW Column Names above).

Accessing Market Aggregates

There are two ways to access market aggregates:

• Making a request to GET https://query.jua.ai/v1/forecast/market-aggregate

• Using POST https://query.jua.ai/v1/forecast/data which allows selecting custom regions such as polygons and bounding boxes

Using the `market-aggregate` endpoint

A convenient GET endpoint that allows fast and easy access to one or multiple market zones.

The Query Engine OpenAPI docs provide an interactive description of all endpoints

Curl

curl -X GET "https://query.jua.ai/v1/forecast/market-aggregate\
?models=ept2\
&init_time=2025-10-22T00:00:00Z\
&weighting=wind_capacity\
&market_zones=DE\
&variables=wind_speed_at_height_level_100m\
&max_prediction_timedelta=72\
&include_time=true \
  -H "X-API-Key: YOUR_API_KEY" \
  -H "Accept: application/json"

Python

import requests

url = "https://query.jua.ai/v1/forecast/market-aggregate"
headers = {
    "X-API-Key": "YOUR_API_KEY",
    "Accept": "application/json"
}
params = {
    "models": "ept2",
    "init_time": "2025-10-22T00:00:00Z",
    "weighting": "wind_capacity",
    "market_zones": ["DE"],  # Germany
    "variables": ["wind_speed_at_height_level_100m"],
    "max_prediction_timedelta": 72, # hours
    "include_time": True,
}

response = requests.get(url, headers=headers, params=params)
data = response.json()

Javascript

const url = new URL("https://query.jua.ai/v1/forecast/market-aggregate");
url.searchParams.append("models", "ept2");
url.searchParams.append("init_time", "2025-10-22T00:00:00Z");
url.searchParams.append("weighting", "wind_capacity");
url.searchParams.append("market_zones", "DE");
url.searchParams.append("variables", "wind_speed_at_height_level_100m");
url.searchParams.append("max_prediction_timedelta", "2880");
url.searchParams.append("include_time", "true");

const response = await fetch(url, {
  method: "GET",
  headers: {
    "X-API-Key": "YOUR_API_KEY",
    Accept: "application/json",
  },
});

const data = await response.json();

Response

{
    "model": [
        "ept2",
        "ept2",
        71 more entries...
    ],
    "init_time": ["2025-10-22T00:00:00Z", ...],
    "prediction_timedelta": [0, 1, ...., 72],
    "time": ["2025-10-22T00:00:00Z", "2025-10-22T01:00:00Z", ...],
    "avg__wind_speed_at_height_level_100m": [0.23, 0.51, ...]
}    

Using the generic `data` endpoint

While adding slightly more overhead to the query, the data endpoint provides more flexibility such as defining custom regions using polygons and bounding boxes. The example below is equivalent to the request to the market-aggregate endpoint above.

Checkout the examples on how to use market aggregates with polygons

Python

import requests

url = "https://query.jua.ai/v1/forecast/data"
headers = {
    "X-API-Key": "YOUR_API_KEY",
    "Content-Type": "application/json",
    "Accept": "application/json"
}
params = {
    "format": "json"
}
payload = {
    "models": ["ept2"],
    "geo": {
        "type": "market_zone",
        "value": ["DE"]
    },
    "variables": ["wind_speed_at_height_level_100m"],
    "init_time": "2025-10-22T00:00:00Z",
    "prediction_timedelta": {
        "start": 0,
        "end": 72
    },
    "weighting": {
        "type": "wind_capacity"
    },
    "group_by": ["model", "init_time", "prediction_timedelta", "time"],
}

response = requests.post(url, headers=headers, params=params, json=payload)
data = response.json()

Javascript

const url = "https://query.jua.ai/v1/forecast/data?format=json";

const payload = {
  models: ["ept2"],
  geo: {
    type: "market_zone",
    value: ["DE"],
  },
  variables: ["wind_speed_at_height_level_100m"],
  init_time: "2025-10-22T00:00:00Z",
  prediction_timedelta: {
    start: 0,
    end: 72,
  },
  weighting: {
    type: "wind_capacity",
  },
  group_by: ["model", "init_time", "prediction_timedelta", "time"],
};

const response = await fetch(url, {
  method: "POST",
  headers: {
    "X-API-Key": "YOUR_API_KEY",
    "Content-Type": "application/json",
    Accept: "application/json",
  },
  body: JSON.stringify(payload),
});

const data = await response.json();

Response

{
    "model": [
        "ept2",
        "ept2",
        71 more entries...
    ],
    "init_time": ["2025-10-08T00:00:00Z", ...],
    "prediction_timedelta": [0, 1, ...., 72],
    "avg__wind_speed_at_height_level_100m": [285.34, 285.91, ...],
}

Countries with Regional Zones

Some countries don't have a single zone code — use their regional zones instead:

Country	Zone Codes
Italy	IT-CNO, IT-CSO, IT-NO, IT-SAR, IT-SIC, IT-SO
Norway	NO-NO1, NO-NO2, NO-NO3, NO-NO4, NO-NO5
Sweden	SE-SE1, SE-SE2, SE-SE3, SE-SE4
Denmark	DK-BHM, DK-DK1, DK-DK2
Japan	JP-CB, JP-CG, JP-HKD, JP-HR, JP-KN, JP-KY, JP-ON, JP-SK, JP-TH, JP-TK

Countries with single zone codes: DE, FR, GB, NL, PL, ES, BE, AT, CH, CZ, PT, GR, IE

DK-BHM (Bornholm) supports unweighted and population-weighted weather queries. Wind-capacity weighting and MW mode are not available for this zone.