Skip to main content
CARLA HQ
DATASET PROFILE / TIME SERIES

Global Energy Forecasting Competition (GEFCom 2014)

GEFCom 2014 (proenfo_gfc14) provides 17,520 hourly observations across a single electrical zone, targeting a 168-hour (7-day ahead) forecast of power grid demand. Conditioned on exogenous air temperature, the benchmark tests medium-term multi-seasonal grid scheduling and dispatch optimization.

Energy & Utilities Operations & Workforce Analytics use casesNACE D35Time Series ForecastingHourly Cadence168 Steps HorizonMarkdown ↗Target: targetfev / AutoGluon ↗
✨ Try in CarlaInstall Chrome Extension ↗100% In-Browser WebGPU • Zero Cloud Upload
17,520Records (Rows)
1Time Series
HourlySampling Cadence
168 HoursForecast Horizon (20 Windows)
24h Diurnal
Dominant Seasonality?Strong 24-hour diurnal cycle combined with a 168-hour weekly profile showing distinct weekend baseload drops, overlaid with non-linear heating/cooling seasonal shifts.
41 Methodsfev Leaderboard (SQL)
BUSINESS CONTEXT & OBJECTIVEPower Generation & Grid Planning (D35)

Business Objective: Global Energy Forecasting Competition (GEFCom 2014)

01Business Context

Transmission system operators (TSOs) and energy trading desks operate on weekly commitment cycles, scheduling baseload and peaker generation 168 hours ahead to ensure grid stability and clear day-ahead power markets.

02Analytical Objective

Forecast hourly zonal electrical load across a full 168-hour ahead horizon using historical consumption telemetry and known future ambient air temperatures.

03Economic & Decision Impact

Under-forecasting risks reserve margin depletion, localized brownouts, and costly spot-market imbalance penalties. Over-forecasting forces unnecessary thermal plant spinning reserves, curtailment of renewable generation, and elevated operating expenditures.

04Temporal Dynamics & Covariate Information Flow

Dominant Seasonality: Strong 24-hour diurnal cycle combined with a 168-hour weekly profile showing distinct weekend baseload drops, overlaid with non-linear heating/cooling seasonal shifts.

Covariates & Information Flow: Future surface air temperature forecasts (`airtemperature`) directly dictate thermo-dependent electric heating and HVAC cooling loads, alongside calendar regressors.

ML Benchmark Narrative

Classical univariate baselines (AutoARIMA, ETS) struggle over the long 168-hour horizon due to phase error accumulation across multi-day diurnal cycles. Feature-engineered LightGBM models with temperature lag interactions excel, while zero-shot foundation models such as Mistral's Timestral and MOIRAI capture complex non-linear load-temperature dynamics without manual feature crafting.

Source Origin:IEEE Power & Energy Society / GEFCom 2014
Creator:Tao Hong et al. (IEEE Working Group on Energy Forecasting) (2016)
License:Open Data / Academic Research
IN-BROWSER FORECAST DEMO

Forecast this dataset in your browser

A representative series is read directly from this dataset’s Parquet file. Move the forecast start, compare held-out observations, and toggle available numeric covariates with Timestral 5M (12 MB) or switch to ci-proto (~346 MB, WebGPU). Benchmark scores above use the stated evaluation setup.

Live sample from dataset Parquet
Timestral 5M · WebGPU
Nord Pool Day-Ahead Electricity PricesHourly · €/MWh
Forecast History
Loading sample data…
Covariates
FEATURE SPECIFICATION & DATA DICTIONARY

Target & Feature Column Definitions

Exhaustive business definitions, measurement units, roles, and target variables across all 4 columns.

INFORMATION AVAILABILITY & TEMPORAL PROTOCOL

Known Future Covariates vs. Historically Observed Covariates

Multi-step forecasting models must respect the arrow of time and real-world information availability. FEV benchmarks and Carla strictly distinguish between features known across the forward prediction window versus realizations available only up to the forecast origin.

FUTURE KNOWN COVARIATESLegitimately available at test time (steps t+1 to t+H)

Scheduled promotions (such as Rossmann's Promo), calendar events (school/state holidays), and operational forecasts (such as Nord Pool's day-ahead grid load and wind power forecasts). These are deterministic or published before market clearing, allowing foundation models to legitimately condition on them across future steps.

HISTORICALLY OBSERVED COVARIATESStrictly restricted to history (time ≤ t)

Stochastic operational observations (such as Rossmann's actual customer footfall Customers, Rohlik's actual precipitation and snowfall, or sudden warehouse shutdowns). Because these are only realized in real-time, future values must never enter a prediction to prevent lookahead target leakage.

Observed Dataset Dynamics: Future surface air temperature forecasts (`airtemperature`) directly dictate thermo-dependent electric heating and HVAC cooling loads, alongside calendar regressors.
ColVariable / Feature NameRoleTypeUnit / ScaleDescription & Business Meaning
CtargetZonal Electrical LoadTARGETNUMMW
Total hourly electric power consumption observed across the regional utility zone.
AidSeries IdentifierSERIES IDCAT—
Unique categorical identifier denoting the electrical zonal load time series.
BtimestampTimestampTIMETIME—
Hourly observation date and time in ISO 8601 format.
DairtemperatureAir TemperatureFUTURE KNOWNNUMdeg F
Observed and forecasted ambient surface air temperature influencing thermal building demand.
• Known future driver: Legitimate lookahead across forecast horizon (steps t+1 to t+H, e.g. calendar/promotion/forecast).
DATASET PREVIEW

Interactive Data Table

Explore rows, feature values, and target labels for Global Energy Forecasting Competition (GEFCom 2014).

Loading dataset...
GLOBAL BENCHMARK PROVENANCE (FEV-BENCH)

Foundation Models vs Traditional ML & Baselines (35 Methods)

Empirical evaluation directly from the AutoGluon fev benchmark suite. Global aggregate ranking: The win rates and skill scores below represent cross-dataset performance across all 100 standardized tasks in fev-bench, rather than results isolated to Global Energy Forecasting Competition (GEFCom 2014).

Benchmark Provenance & Latency Measurement:
100-Task Global Aggregation:The rankings evaluate zero-shot foundation models across 100 heterogeneous benchmark tasks in fev-bench, measuring aggregate generalization across diverse frequencies and prediction horizons.
FEV Python vs. In-Browser WebGPU Timing:The "FEV Latency" column reports server-side Python execution times from the official AutoGluon FEV benchmark. Independently, Carla measures and reports true client-side WebGPU execution inside Google Sheets (62ms on this dataset) with zero server reliance.
#1 Official Ranked Model?ci-proto from Continuous Intelligence ranks #1 on the 35-method official fev-bench view (90.0% SQL win rate; #3 overall at 84.37% across all 52 evaluated models) and can also run locally in our browser WebGPU demo.
ci-proto90.0% Win Rate • 0.53s FEV latency
Evaluated Methods?35 official leaderboard methods (52 total evaluated including closed APIs and checkpoint variants across foundation models, deep neural architectures, trees, and statistical baselines).
35 Official52 total • Continuous Intelligence, H2O.ai, Mistral AI, Google, AWS
Primary Evaluation Metric?Scaled Quantile Loss (SQL) penalizes both point prediction errors and quantile miscalibration across 9 probabilistic levels (p10 to p90).
SQL & MASEProbabilistic calibration
Independent Browser WebGPU?Carla Engine executes Timestral-123M locally in Google Sheets via WebGPU shaders, while ci-proto can also be tested directly in-browser via ONNX Runtime WebGPU and Rust/WASM above.
62ms / windowTimestral-123M (Carla Engine) • Zero server upload
RankMethod / ModelFamilyOrganization
SQL Win Rate?Average win rate (%) against all competing models across the benchmark tasks under Scaled Quantile Loss.
SQL Skill?Normalized skill score relative to the Seasonal Naive baseline (0% = baseline, higher is better).
MASE Win Rate?Average win rate (%) under Mean Absolute Scaled Error.
FEV Latency?FEV benchmark reference latency measured in server Python (median seconds per 100 series). Independent in-browser WebGPU execution for Carla is reported in the telemetry card above.
Link
#1ci-protoFoundation ModelContinuous Intelligence90.0%+51.5%88.4%0.53sRepo ↗
#2TimeH2O-v1Foundation ModelH2O.ai89.7%+50.8%88.2%2.82sRepo ↗
#3Timestral-209MFoundation ModelMistral AI87.1%+50.4%85.4%0.51sRepo ↗
#4t0-gammaFoundation ModelThe Forecasting Company84.7%+49.4%82.0%4.38sRepo ↗
#5Simons-297MFoundation ModelAbel AI Lab82.0%+48.4%80.9%0.94sRepo ↗
#6timesfm-3Foundation ModelGoogle79.3%+49.0%77.6%3.41sRepo ↗
#7iAmTime-base-forecastFoundation ModelWalmart Labs73.7%+46.8%72.5%0.56sRepo ↗
#8Chronos-2Foundation ModelAWS72.2%+47.3%69.9%0.84sRepo ↗
#9t0-betaFoundation ModelThe Forecasting Company69.9%+46.6%66.9%2.29sRepo ↗
#10Toto-2.0-2.5BBaselineCommunity69.4%+44.4%67.8%5.58sRepo ↗
#11TiRex-2Foundation ModelNX-AI69.2%+45.5%66.3%0.27sRepo ↗
#12TiRexFoundation ModelNX-AI62.2%+42.6%58.0%0.24sRepo ↗
#13Toto-2.0-22mBaselineCommunity61.6%+42.9%59.4%0.25sRepo ↗
#14TimesFM-2.5BaselineCommunity58.4%+42.2%56.7%1.87sRepo ↗
#15TabPFN-TS-3Foundation ModelPrior Labs55.5%+43.1%53.3%235sRepo ↗
#16TS-ICLFoundation ModelEDF-Lab55.1%+43.1%53.4%2.49sRepo ↗
#17citras-fmFoundation ModelHitachi52.9%+41.2%50.6%0.60sRepo ↗
#18FlowStateFoundation ModelIBM52.7%+39.9%53.0%2.29sRepo ↗
#19Toto-1.0BaselineCommunity51.6%+40.8%48.8%22.35sRepo ↗
#20TabPFN-TSFoundation ModelPrior Labs50.5%+41.5%48.0%109sRepo ↗
#21Moirai-2.0BaselineCommunity46.2%+39.3%44.2%0.35sRepo ↗
#22Chronos-BoltFoundation ModelAWS45.4%+38.9%42.9%0.26sRepo ↗
#23TFTDeep Learning—34.7%+32.2%31.6%829sRepo ↗
#24Sundial-BaseFoundation ModelTsinghua University30.8%+33.4%37.0%8.01sRepo ↗
#25Stat. EnsembleBaselineCommunity30.6%+20.7%33.6%147sRepo ↗
#26PatchTSTDeep Learning—30.0%+30.1%27.8%664sRepo ↗
#27DeepARDeep Learning—28.9%+29.1%25.9%1113sRepo ↗
#28AutoARIMAStatistical Baseline—26.6%+21.3%24.2%20.14sRepo ↗
#29AutoETSStatistical Baseline—22.5%-26.8%24.1%3.48sRepo ↗
#30CatBoostTree Ensemble—21.8%+23.0%37.5%29.48sRepo ↗
#31LightGBMTree Ensemble—20.3%+21.0%35.7%3.03sRepo ↗
#32AutoThetaStatistical Baseline—17.9%+5.5%22.2%3.28sRepo ↗
#33Seasonal NaiveBaselineCommunity12.0%0.0%13.2%0.48sRepo ↗
#34NaiveStatistical Baseline—7.8%-45.4%12.2%0.47sRepo ↗
#35DriftStatistical Baseline—7.0%-45.8%11.1%0.45sRepo ↗
STATISTICAL DISTRIBUTIONS

Variable Schema & Summary Distributions

Observed numerical ranges, category cardinalities, missing rates, and sample values across 17,520 rows.

ColVariable NameTypeMissingDistinctSummary Stats / DistributionSample Values
CtargetTARGETNUM02,502[2107, 5224] μ=3328.2 σ=561.3286627102605
AidCAT01—loadloadload
BtimestampCAT017,520—2013-01-01T01:00:002013-01-01T02:00:002013-01-01T03:00:00
DairtemperatureFUTURE KNOWNNUM0311[-13.670000076293945, 95] μ=46.8 σ=19.829.32999992370605529.32999992370605530
FREQUENTLY ASKED QUESTIONS

Frequently Asked Questions: Global Energy Forecasting Competition (GEFCom 2014)

Common questions regarding the Global Energy Forecasting Competition (GEFCom 2014) dataset, machine learning task formulations, and in-browser tabular inference.

What is the forecasting horizon for GEFCom 2014 (proenfo_gfc14)?

The horizon is 168 hours (7 full calendar days), reflecting standard transmission grid scheduling and weekly generation dispatch cycles.

Can I run zero-server forecasts on this dataset in Google Sheets?

Yes, Carla enables zero-server in-context time series forecasting directly in spreadsheets using Mistral's Timestral foundation model without moving data outside your sheet.

Which models perform best on GEFCom 2014?

Modern zero-shot time series foundation models like Mistral's Timestral and tuned gradient-boosted trees (LightGBM) with temperature interaction terms outperform classical statistical baselines like AutoARIMA and Exponential Smoothing.

LIVE EVALUATION IN GOOGLE SHEETS

Forecast with Carla: Global Energy Forecasting Competition (GEFCom 2014)

Open the pre-loaded Google Sheet and let Carla generate multi-step probabilistic forecasts using Mistral's Timestral directly inside your browser with zero server uploads.

Step 1

Launch Carla

Click once to open the spreadsheet and Carla side panel together.

Step 2

Review the Setup

Carla auto-detects the target sequence, frequency cadence, and temporal covariates.

Step 3

Forecast & Calibrate

Generate multi-step quantile forecasts with zero server uploads.

Provenance & Attribution

Hong, T., Pinson, P., Fan, S., Zareipour, H., Troccoli, A., & Hyndman, R. J. (2016). Probabilistic energy forecasting: Global Energy Forecasting Competition 2014 and beyond. International Journal of Forecasting, 32(3), 896-913.

License: Open AccessData Source: autogluon/fev_datasetsOriginal Source: https://doi.org/10.1016/j.ijforecast.2016.02.001