{"id":553,"date":"2026-06-09T05:17:40","date_gmt":"2026-06-09T05:17:40","guid":{"rendered":"https:\/\/jua.ai\/articles\/historical-weather-data-europe-2026\/"},"modified":"2026-06-09T05:17:40","modified_gmt":"2026-06-09T05:17:40","slug":"historical-weather-data-europe-2026","status":"publish","type":"post","link":"https:\/\/jua.ai\/articles\/historical-weather-data-europe-2026\/","title":{"rendered":"Historical Weather Data Europe: 2026 Directory for Analysts"},"content":{"rendered":"<p><em>Written by: Olivier Lam, Physical AI Team, Jua.ai AG<\/em><\/p>\n<h2 id=\"key-takeaways\">Key Takeaways for European Weather Backtests<\/h2>\n<ul>\n<li>Five authoritative sources supply European historical weather data in 2026: ERA5, E-OBS, Meteostat, Open-Meteo Historical API, and meteoblue history+.<\/li>\n<li>ERA5 provides the longest multi-decade gridded record (1990\u2013present) at 0.25\u00b0 resolution and is the default reference for wind and solar backtests.<\/li>\n<li>E-OBS and Meteostat deliver station-derived observational ground truth ideal for validating near-surface temperature and precipitation forecasts.<\/li>\n<li>Open-Meteo\u2019s CERRA integration offers 5.5 km European resolution (1984\u20132021), while meteoblue history+ supplies commercial point-location archives from 1940 onward.<\/li>\n<li>Jua for Energy turns any of these datasets into model-evaluation deliverables across 25+ models in roughly five minutes, so you can <a href=\"https:\/\/meetings-eu1.hubspot.com\/guett\/energy-trading?uuid=d780665f-ff71-439c-addf-c80e49af0627\" target=\"_blank\">book a demo<\/a> and see it in practice.<\/li>\n<\/ul>\n<h2>Core Sources for European Historical Weather Data<\/h2>\n<p>European historical weather data lives in public reanalysis archives, station observation networks, and commercial APIs. No single source covers every use case. Reanalysis products such as ERA5 offer multi-decade spatial consistency, but they trade off station-level precision. Station-interpolated grids such as E-OBS preserve observational fidelity, yet they introduce interpolation uncertainty in data-sparse regions. API-first services such as Open-Meteo and meteoblue history+ cut pipeline work, but they reduce flexibility compared with raw-format archives. The five sources below form a practical working set for quant developers, meteorologists, and energy analysts in European power and gas markets. Once you have picked a reference dataset, you can validate your own forecasts against it on the Jua platform and see how your models stack up across 25+ alternatives.<\/p>\n<p><a href=\"https:\/\/athena.jua.ai\" target=\"_blank\"><strong>Test these datasets against live forecasts on the Jua platform. Run benchmarks on your own region and variables at athena.jua.ai.<\/strong><\/a><\/p>\n<h2>Copernicus ERA5 for Multi\u2011Decade Gridded Records<\/h2>\n<p>ERA5 is ECMWF&#8217;s fifth-generation global atmospheric reanalysis. It covers 1990 to present at 0.25\u00b0 \u00d7 0.25\u00b0 horizontal resolution for atmospheric fields, with <a href=\"https:\/\/cds.climate.copernicus.eu\/datasets\/reanalysis-era5-land\" target=\"_blank\" rel=\"noindex nofollow\">ERA5-Land extending coverage to 0.1\u00b0 (~9 km native) from January 1950<\/a>. Key variables include 2 m temperature, 10 m and 100 m wind speed and direction, surface solar radiation downwards (SSRD), total precipitation, sea-level pressure, and soil moisture at four depth layers. Data is distributed via the <a href=\"https:\/\/cds.climate.copernicus.eu\" target=\"_blank\" rel=\"noindex nofollow\">Copernicus Climate Data Store (CDS)<\/a> in GRIB format and as Analysis Ready Cloud Optimized (ARCO) Zarr for efficient time-series retrieval.<\/p>\n<p>ERA5 is updated daily with approximately five days of latency. A preliminary ERA5T release appears sooner and may differ from the final version released two to three months later. On 25 February 2026, ECMWF updated its area-extraction software, so sub-area requests now return only grid points inside the requested bounding box, not values interpolated onto a new grid. Teams that rely on prior coordinate grids should validate existing pipelines against fresh downloads. For energy trading, ERA5&#8217;s 100 m wind variable and SSRD record make it the standard reference for wind and solar generation backtests across multi-decade horizons.<\/p>\n<h2>E-OBS for Gridded Station-Based Ground Truth<\/h2>\n<p><a href=\"https:\/\/www.ecad.eu\/download\/ensembles\/download.php\" target=\"_blank\" rel=\"noindex nofollow\">E-OBS<\/a> is a daily gridded observational dataset for Europe produced by the European Climate Assessment &amp; Dataset (ECA&amp;D) project. It covers Europe from 1950 to near-present on a 0.1\u00b0 regular grid, derived by interpolating quality-controlled station observations from the ECA&amp;D network. Core variables are daily maximum and minimum temperature, daily mean temperature, precipitation sum, sea-level pressure, relative humidity, and global radiation. Data is distributed as NetCDF files via the <a href=\"https:\/\/surfobs.climate.copernicus.eu\/dataaccess\/access_eobs.php\" target=\"_blank\" rel=\"noindex nofollow\">Copernicus Climate Change Service<\/a>. Because E-OBS is built directly from surface station records rather than model assimilation, it is the preferred reference for validating near-surface temperature and precipitation forecasts against observed climatology, especially for demand-side energy models where 2 m temperature drives load forecasts.<\/p>\n<h2>Meteostat for Direct Station Observations<\/h2>\n<p><a href=\"https:\/\/meteostat.net\" target=\"_blank\" rel=\"noindex nofollow\">Meteostat<\/a> aggregates historical weather observations from public sources including NOAA, DWD, and national meteorological services. It covers thousands of stations worldwide with long-term time series to present. Variables include temperature, precipitation, wind speed and direction, pressure, and sunshine duration at daily and hourly resolution where station records permit. Data is accessible via the <a href=\"https:\/\/dev.meteostat.net\/python\/\" target=\"_blank\" rel=\"noindex nofollow\">Meteostat Python library<\/a> and a JSON API. For energy analysts, Meteostat is most useful for point-location validation of temperature and wind forecasts against actual station observations, particularly in markets where a specific plant or grid node is the unit of analysis.<\/p>\n<pre><code># Copy-paste: retrieve hourly station data with Meteostat from datetime import datetime from meteostat import Point, Hourly start = datetime(2023, 1, 1) end = datetime(2023, 12, 31) # Frankfurt am Main, Germany location = Point(50.11, 8.68, 112) data = Hourly(location, start, end) df = data.fetch() print(df.head()) <\/code><\/pre>\n<h2>Open-Meteo Historical API for Zero-Friction Reanalysis Access<\/h2>\n<p>Open-Meteo&#8217;s Historical Weather API combines several reanalysis sources behind a single REST endpoint: ERA5 and ERA5-Land at their standard resolutions and coverage periods, plus ECMWF IFS at 9 km from 2017. For European-specific work, CERRA high-resolution reanalysis provides 5.5 km coverage over Europe from 1 September 1984 to 30 June 2021 and is included in the Best Match model selection. The API exposes a range of hourly variables including 2 m temperature, wind speed at 10 m and 100 m, shortwave radiation, precipitation, soil temperature and moisture, and sea-level pressure. Responses are returned as JSON, CSV, or XLSX. Non-commercial use is free below 10,000 daily calls, while commercial use requires an API key. For energy model validation, the 100 m wind variable and CERRA&#8217;s 5.5 km European resolution make Open-Meteo a practical low-friction entry point for wind-generation backtests.<\/p>\n<pre><code># Copy-paste: fetch hourly 100m wind and solar radiation via Open-Meteo import requests params = { \"latitude\": 52.52, \"longitude\": 13.41, \"start_date\": \"2023-01-01\", \"end_date\": \"2023-12-31\", \"hourly\": \"wind_speed_100m,shortwave_radiation\", \"models\": \"era5\", \"timezone\": \"Europe\/Berlin\", } response = requests.get(\"https:\/\/archive-api.open-meteo.com\/v1\/archive\", params=params) data = response.json() print(data[\"hourly\"].keys()) <\/code><\/pre>\n<h2>meteoblue history+ for Commercial Point Locations<\/h2>\n<p><a href=\"https:\/\/www.meteoblue.com\/en\/weather\/archive\/export\" target=\"_blank\" rel=\"noindex nofollow\">meteoblue history+<\/a> is a commercial point-location historical weather archive covering Europe and the globe from 1940 to present. It is derived from the meteoblue NEMS reanalysis and blended with station observations. Key variables include temperature, wind speed and direction at multiple heights, solar radiation, precipitation, humidity, and pressure. Data is accessible via a REST API and CSV export, with commercial licensing required. For energy trading applications, meteoblue history+ is relevant when you need high-resolution point-location data for a specific turbine site or solar farm without processing gridded reanalysis files.<\/p>\n<h2>Side-by-Side Comparison of Key Datasets<\/h2>\n<table>\n<thead>\n<tr>\n<th>Source<\/th>\n<th>Spatial Resolution \/ Coverage<\/th>\n<th>Access Method &amp; Format<\/th>\n<th>Cost &amp; Update Cadence<\/th>\n<\/tr>\n<\/thead>\n<tbody>\n<tr>\n<td>ERA5 \/ ERA5-Land<\/td>\n<td>0.25\u00b0 global (ERA5); 0.1\u00b0 global (ERA5-Land); 1990\u2013present (ERA5), 1950\u2013present (ERA5-Land). Variables: wind 10 m &amp; 100 m, SSRD, temperature, precipitation, soil moisture.<\/td>\n<td>CDS API; GRIB and ARCO Zarr<\/td>\n<td>Free (Copernicus); daily updates, ~5-day latency<\/td>\n<\/tr>\n<tr>\n<td><a href=\"https:\/\/www.ecad.eu\/download\/ensembles\/download.php\" target=\"_blank\" rel=\"noindex nofollow\">E-OBS<\/a><\/td>\n<td>0.1\u00b0 Europe; 1950\u2013near-present. Variables: daily max\/min\/mean temperature, precipitation, pressure, humidity, radiation.<\/td>\n<td>NetCDF download via Copernicus C3S<\/td>\n<td>Free; periodic version releases<\/td>\n<\/tr>\n<tr>\n<td><a href=\"https:\/\/meteostat.net\" target=\"_blank\" rel=\"noindex nofollow\">Meteostat<\/a><\/td>\n<td>Point stations worldwide; long-term time series to present. Variables: temperature, wind, precipitation, pressure, sunshine.<\/td>\n<td>Python library; JSON API<\/td>\n<td>Free (open-source); near-real-time station updates<\/td>\n<\/tr>\n<tr>\n<td>Open-Meteo Historical API<\/td>\n<td>0.25\u00b0 (ERA5, 1990\u2013); 0.1\u00b0 (ERA5-Land, 1950\u2013); 5.5 km Europe (CERRA, 1984\u20132021). Variables: wind 10 m &amp; 100 m, SSRD, temperature, precipitation, soil layers, and others.<\/td>\n<td>REST \/v1\/archive; JSON, CSV, XLSX<\/td>\n<td>Free &lt;10,000 calls\/day; commercial API key for higher volume; near-real-time<\/td>\n<\/tr>\n<tr>\n<td><a href=\"https:\/\/www.meteoblue.com\/en\/weather\/archive\/export\" target=\"_blank\" rel=\"noindex nofollow\">meteoblue history+<\/a><\/td>\n<td>Point-location global; 1940\u2013present. Variables: temperature, wind at multiple heights, solar radiation, precipitation, humidity.<\/td>\n<td>REST API; CSV export<\/td>\n<td>Commercial license required; near-real-time<\/td>\n<\/tr>\n<\/tbody>\n<\/table>\n<h2>Using Historical Data for Energy Model Validation<\/h2>\n<p>Model validation in energy trading quantifies forecast skill, meaning how closely a model&#8217;s predicted wind speed, solar irradiance, or temperature matches what actually happened. The standard workflow compares model hindcasts, which are retrospective forecasts generated with the same model architecture and initialization procedure used operationally, against a ground-truth reference. ERA5 serves as the most common gridded reference for long-horizon backtests. E-OBS and Meteostat station records serve as observational ground truth for near-surface variables. Forecast skill metrics such as RMSE (root mean square error) and CRPS (continuous ranked probability score, a proper scoring rule for probabilistic forecasts) are computed across lead times, seasons, and geographic zones to show where a model adds edge and where it degrades.<\/p>\n<p>The practical bottleneck usually comes from engineering overhead, not from access to reference data. Teams must run the same backtest across multiple models, normalize schemas, and produce a comparable skill table. <a href=\"https:\/\/jua.ai\/articles\/ai-powered-energy-analytics\" target=\"_blank\">Jua for Energy, the first applied product from Jua, a foundation model and agent company, provides hindcast data and runs the same backtest across 25+ models<\/a> via Athena, the AI agent instrumented with the Jua for Energy tool surface, or directly through the Python SDK (<code>pip install jua<\/code>). The 25-model set includes 10 proprietary AI models from the EPT family, including EPT-2e, the ensemble variant that updates four times per day and supports native forecasting down to 5 km resolution, plus 15 third-party NWP and AI models including ECMWF HRES, ECMWF ENS, ECMWF AIFS, NOAA GFS, Microsoft Aurora, and GFS GraphCast. EPT-2e beats the 50-member ECMWF ENS mean on both RMSE and CRPS at virtually every lead time, as documented in the peer-reviewed technical report at <a href=\"https:\/\/arxiv.org\/abs\/2507.09703\" target=\"_blank\" rel=\"noindex nofollow\">arXiv:2507.09703<\/a>. <a href=\"https:\/\/www.bloomberg.com\/news\/articles\/2026-03-26\/energy-traders-turn-to-ai-to-forecast-the-weather-forecast?embedded-checkout=true\" target=\"_blank\">European energy traders are increasingly using AI tools to evaluate forecast model revisions<\/a>, and a platform that surfaces those comparisons automatically, rather than requiring manual pipeline assembly, lets teams act before the market instead of reacting after it.<\/p>\n<p><a href=\"https:\/\/athena.jua.ai\" target=\"_blank\"><strong>Skip manual pipeline work and see cross-model benchmarks in under 5 minutes at athena.jua.ai.<\/strong><\/a><\/p>\n<h2>Choosing the Right Dataset for Your Use Case<\/h2>\n<table>\n<thead>\n<tr>\n<th>Use Case<\/th>\n<th>Recommended Source<\/th>\n<th>Rationale<\/th>\n<\/tr>\n<\/thead>\n<tbody>\n<tr>\n<td>Multi-decade wind generation backtest (continental Europe)<\/td>\n<td>ERA5 (100 m wind, 1990\u2013present, 0.25\u00b0)<\/td>\n<td>Longest consistent record. The 100 m wind variable maps to turbine hub heights. ARCO Zarr format supports large-payload retrieval.<\/td>\n<\/tr>\n<tr>\n<td>High-resolution wind ramp studies (Europe, 1984\u20132021)<\/td>\n<td>Open-Meteo \/ CERRA (5.5 km, 1984\u20132021)<\/td>\n<td>CERRA&#8217;s 5.5 km European resolution resolves mesoscale ramp events that 0.25\u00b0 ERA5 smooths. The REST endpoint reduces pipeline overhead.<\/td>\n<\/tr>\n<tr>\n<td>Near-surface temperature validation for load forecasting<\/td>\n<td><a href=\"https:\/\/www.ecad.eu\/download\/ensembles\/download.php\" target=\"_blank\" rel=\"noindex nofollow\">E-OBS<\/a> or <a href=\"https:\/\/meteostat.net\" target=\"_blank\" rel=\"noindex nofollow\">Meteostat<\/a><\/td>\n<td>Station-derived observations provide ground truth unaffected by reanalysis model bias. E-OBS supports gridded spatial coverage, while Meteostat supports specific node-level validation.<\/td>\n<\/tr>\n<tr>\n<td>Solar irradiance backtest for PV generation model<\/td>\n<td>ERA5 (SSRD, 1990\u2013present)<\/td>\n<td>ERA5 SSRD is the only multi-decade gridded solar radiation record with consistent methodology. ERA5-Land at 0.1\u00b0 adds higher spatial detail.<\/td>\n<\/tr>\n<tr>\n<td>Point-location commercial archive (specific plant site)<\/td>\n<td><a href=\"https:\/\/www.meteoblue.com\/en\/weather\/archive\/export\" target=\"_blank\" rel=\"noindex nofollow\">meteoblue history+<\/a><\/td>\n<td>Blended reanalysis and station data at point location. REST and CSV delivery avoid gridded file processing.<\/td>\n<\/tr>\n<tr>\n<td>Cross-model forecast skill evaluation (25+ models, any region)<\/td>\n<td>Jua for Energy hindcast via <a href=\"https:\/\/athena.jua.ai\" target=\"_blank\">Athena<\/a> or <code>pip install jua<\/code><\/td>\n<td>Single schema across the EPT family and third-party NWP. Backtests complete in minutes, with no pipeline assembly required.<\/td>\n<\/tr>\n<\/tbody>\n<\/table>\n<h2>Conclusion: From Raw Weather Data to Trading-Ready Benchmarks<\/h2>\n<p>The five sources in this directory, ERA5 and ERA5-Land, E-OBS, Meteostat, Open-Meteo Historical API, and meteoblue history+, cover the practical range of European historical weather data needs for quant developers and energy analysts. ERA5 remains the default for multi-decade gridded backtests. E-OBS and Meteostat provide observational ground truth for near-surface validation. Open-Meteo&#8217;s CERRA integration offers 5.5 km spatial resolution over Europe from 1984 to 2021. meteoblue history+ serves commercial point-location requirements.<\/p>\n<p>Accessing the data is only the first step. Turning it into model-evaluation deliverables such as skill scores across 25+ models, ensemble calibration curves, and wind-ramp detection rates by lead time consumes most engineering time on many teams. <a href=\"https:\/\/nebius.com\/customer-stories\/jua\" target=\"_blank\">Jua for Energy, built on EPT-2 and the Athena agent, delivers hindcast-backed benchmarks across 25+ models<\/a> with ERA5 as the reference baseline and EPT-2e as the ensemble benchmark that beats the 50-member ECMWF ENS mean on RMSE and CRPS at virtually every lead time. A 1 GW wind portfolio that gains four percentage points of forecast accuracy saves approximately \u20ac1.5 M per year. A 1 GW solar portfolio at the same accuracy gain saves approximately \u20ac3 M per year. The benchmark provides the proof.<\/p>\n<p><a href=\"https:\/\/athena.jua.ai\" target=\"_blank\"><strong>Get started with head-to-head model comparisons across 25+ forecasts at athena.jua.ai.<\/strong><\/a><\/p>\n<h2>Frequently Asked Questions<\/h2>\n<h3>What is the best free source of historical weather data for Europe?<\/h3>\n<p>ERA5, distributed via the Copernicus Climate Data Store, is the most widely used free source for European historical weather data. It covers 1990 to present at 0.25\u00b0 global resolution, with ERA5-Land extending to 0.1\u00b0 from 1950. Both are free under the Copernicus licence. For a low-friction REST alternative, Open-Meteo&#8217;s Historical Weather API combines ERA5, ERA5-Land, and ECMWF IFS behind a single endpoint and is free for non-commercial use below 10,000 daily calls. For station-level observations, Meteostat provides an open-source Python library with records from thousands of stations worldwide. The right choice depends on whether the use case requires spatial consistency across decades, high spatial resolution over Europe via CERRA, or observational ground truth at specific locations through Meteostat or E-OBS.<\/p>\n<h3>How do quant developers and energy analysts use historical weather data for backtesting?<\/h3>\n<p>Backtesting in energy trading compares a model&#8217;s historical forecast outputs, or hindcasts, against a reference dataset such as ERA5 or station observations to compute forecast skill metrics including RMSE and CRPS across lead times, seasons, and geographic zones. The workflow highlights where a model adds edge, for example whether a wind model outperforms ECMWF HRES on 100 m wind speed at 48-hour lead time over northern Germany in winter. The same hindcast data can backtest systematic trading strategies, such as whether a wind-ramp detection signal based on EPT-2e ensemble spread would have generated positive expected value over the prior two winters. Jua for Energy provides hindcast data across multiple models via the Python SDK and runs full backtests through Athena with the speed described earlier, which removes the pipeline assembly that typically consumes a large share of engineering time.<\/p>\n<h3>What variables in European historical weather datasets are most relevant for energy trading?<\/h3>\n<p>The variables that drive the largest share of European power and gas P&amp;L are 100 m wind speed and direction for wind generation, surface solar radiation downwards or SSRD for solar generation, 2 m temperature for heating and cooling demand and gas consumption, and total precipitation for hydro reservoir inflows. ERA5 and ERA5-Land cover all four with multi-decade consistency. Open-Meteo&#8217;s CERRA layer adds 5.5 km spatial resolution for wind over Europe from 1984 to 2021. For load forecasting specifically, 2 m temperature from E-OBS or Meteostat station records provides observational ground truth that reanalysis products can underestimate in urban heat-island zones. EPT-2, the deterministic flagship in Jua for Energy, outperforms ECMWF HRES on all four of these variables, 10 m wind, 100 m wind, 2 m temperature, and SSRD, across the full 0\u2013240 hour lead-time range, as documented in the peer-reviewed technical report at arXiv:2507.09703.<\/p>\n<h3>What changed in ERA5 data access in early 2026?<\/h3>\n<p>On 25 February 2026, ECMWF updated the software used for area extraction from ERA5 single levels, ERA5 pressure levels, ERA5 monthly means, ERA5 daily statistics, ERA5-Land hourly, ERA5-Land monthly means, ERA5-Land daily statistics, and the complete ERA5 global atmospheric reanalysis. The change stopped interpolating data onto a new grid when a user-supplied bounding box did not exactly match existing grid coordinates. The system now returns only the original grid points inside the requested bounding box. As a result, requests for a geographical area may return data on a grid displaced by up to one grid length from the prior output, and the number of returned grid points may differ. Point-location requests on affected datasets now fail because no grid points lie inside a zero-area bounding box, and ECMWF directs those users to the post-processed ERA5 hourly time-series and ERA5-Land hourly time-series datasets. Teams with existing ERA5 sub-area pipelines should download a fresh sample and compare it with prior files to determine whether workflow modifications are required. CDS server-side caching can retain pre-upgrade coordinate grids, and adding a unique parameter to the API request forces retrieval of the post-upgrade grid.<\/p>\n<h3>How does Jua for Energy differ from simply downloading ERA5 and running my own benchmarks?<\/h3>\n<p>Downloading ERA5 and building a benchmarking harness from scratch requires ingesting GRIB or Zarr files, normalizing schemas across multiple models, implementing RMSE and CRPS scoring, and maintaining the pipeline as model versions and data formats change. That work typically takes a quant team a quarter to stand up and ongoing effort to maintain. Jua for Energy provides hindcast data across 25+ models, including EPT-2e, ECMWF HRES, ECMWF ENS, ECMWF AIFS, Microsoft Aurora, and GFS GraphCast, through a single REST API and Python SDK with a unified schema and Apache Arrow support for large payloads. Athena, the AI agent instrumented with the Jua for Energy tool surface, runs a full cross-model backtest from a natural-language query with the speed described earlier. ERA5 is available as the reference baseline within the platform. The result matches the skill evaluation a team would build manually, but it arrives in minutes instead of months and removes pipeline maintenance overhead.<\/p>\n","protected":false},"excerpt":{"rendered":"<p>Explore top European historical weather data sources for backtests. Jua turns ERA5, E-OBS &amp; more into model insights in minutes. Start today.<\/p>\n","protected":false},"author":103,"featured_media":552,"comment_status":"open","ping_status":"open","sticky":false,"template":"","format":"standard","meta":{"inline_featured_image":false,"footnotes":""},"categories":[11],"tags":[],"class_list":["post-553","post","type-post","status-publish","format-standard","has-post-thumbnail","hentry","category-weather-forecasting"],"_links":{"self":[{"href":"https:\/\/jua.ai\/articles\/wp-json\/wp\/v2\/posts\/553","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/jua.ai\/articles\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/jua.ai\/articles\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/jua.ai\/articles\/wp-json\/wp\/v2\/users\/103"}],"replies":[{"embeddable":true,"href":"https:\/\/jua.ai\/articles\/wp-json\/wp\/v2\/comments?post=553"}],"version-history":[{"count":0,"href":"https:\/\/jua.ai\/articles\/wp-json\/wp\/v2\/posts\/553\/revisions"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/jua.ai\/articles\/wp-json\/wp\/v2\/media\/552"}],"wp:attachment":[{"href":"https:\/\/jua.ai\/articles\/wp-json\/wp\/v2\/media?parent=553"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/jua.ai\/articles\/wp-json\/wp\/v2\/categories?post=553"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/jua.ai\/articles\/wp-json\/wp\/v2\/tags?post=553"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}