Upload germany-electricity-demand-forecaster model

README.md

@@ -0,0 +1,148 @@
+---
+language:
+- en
+- de
+license: apache-2.0
+tags:
+- time-series-forecasting
+- energy-forecasting
+- electricity-demand
+- day-ahead-prices
+- lightgbm
+- tabular-regression
+- conformal-prediction
+- germany
+- europe
+- entsoe
+pipeline_tag: tabular-regression
+datasets:
+- hsilvosa/entsoe-day-ahead
+metrics:
+- mae
+- rmse
+- pinball_loss
+- winkler_score
+model-index:
+- name: germany-demand-forecaster
+  results:
+  - task:
+      type: tabular-regression
+      name: Electricity Demand Forecasting
+    dataset:
+      name: ENTSO-E Germany Bidding Zone (DE)
+      type: hsilvosa/entsoe-day-ahead
+    metrics:
+    - name: MAE
+      type: mae
+      value: 440.218
+    - name: RMSE
+      type: rmse
+      value: 663.288
+    - name: Empirical Interval Coverage (80% Nominal)
+      type: coverage
+      value: 76.5%
+---
+# Electricity Demand Forecaster for Germany (DE)
+
+High-accuracy calibrated quantile LightGBM model for forecasting Germany **demand**.
+Resolution: 15-minute intervals. Trained on multi-year data (2023–2026) from **ENTSO-E**,
+featuring multi-scale lags, cyclical encodings, and **conformal calibration**
+for well-calibrated 80% prediction intervals ($P10, P50, P90$).
+
+## Model Highlights
+
+- **Country / Zone**: Germany (`DE`)
+- **Target**: Electricity Demand in `MW`
+- **Resolution**: 15-minute intervals
+- **Outputs**: Point forecast ($P50$), 80% prediction interval ($P10$ to $P90$)
+- **Algorithm**: LightGBM Multi-Quantile Regressor with Conformal Calibration & Monotonicity
+- **Dataset**: ENTSO-E European Transparency Platform (Zone: `DE`)
+- **Training Samples**: 99,380 observations (2023–2026)
+
+## Performance & Benchmark Comparison
+
+Evaluated on out-of-sample test sets against official seasonal persistence benchmarks:
+
+| Metric | LightGBM Forecaster | 7-Day Seasonal Persistence | Improvement |
+|---|---:|---:|---:|
+| **MAE** | **440.218 MW** | 5542.114 MW | **+92.1%** |
+| **RMSE** | **663.288 MW** | — | — |
+- **WAPE**: 0.80%
+| **P10 Pinball Loss** | 237.595 | — | — |
+| **P90 Pinball Loss** | 155.508 | — | — |
+| **P10–P90 Interval Coverage** | **76.5%** | — | Target: 75–85% |
+| **Winkler Score** | 3931.027 | — | — |
+
+## Quickstart: Python Inference
+
+```python
+import pandas as pd
+from huggingface_hub import hf_hub_download
+import joblib
+
+# 1. Download model artifacts
+model_path = hf_hub_download(repo_id="ORGANIZATION/germany-demand-forecaster", filename="models.joblib")
+models = joblib.load(model_path)
+
+# 2. Predict P10, P50 (point), and P90 quantiles
+X_test = pd.read_csv("sample_input.csv")
+p10 = models[0.1].predict(X_test)
+p50 = models[0.5].predict(X_test)
+p90 = models[0.9].predict(X_test)
+
+print("Forecast Point Estimate:", p50[:5])
+print("80% Lower Bound (P10):", p10[:5])
+print("80% Upper Bound (P90):", p90[:5])
+```
+
+## Features Used
+
+The model uses 37 leakage-safe features:
+- `hour`
+- `quarter`
+- `day_of_week`
+- `day_of_year`
+- `month`
+- `is_weekend`
+- `is_holiday`
+- `is_morning_peak`
+- `is_evening_peak`
+- `sin_hour`
+- `cos_hour`
+- `sin_day_of_week`
+- `cos_day_of_week`
+- `sin_day_of_year`
+- `cos_day_of_year`
+- `lag_1h`
+- `lag_2h`
+- `lag_3h`
+- `lag_4h`
+- `lag_24h`
+- `lag_48h`
+- `lag_7d`
+- `lag_14d`
+- `diff_1h`
+- `diff_2h`
+- `diff_24h`
+- `diff_7d`
+- `acceleration_1h`
+- `ema_4step`
+- `ema_12step`
+- `rolling_std_4step`
+- `rolling_mean_24h`
+- `rolling_std_24h`
+- `rolling_min_24h`
+- `rolling_max_24h`
+- `rolling_mean_7d`
+- `rolling_std_7d`
+
+## Intended Use & Advisory
+
+This model is intended for research, energy market analytics, grid load planning,
+and educational forecasting demonstrations. It is advisory only and not intended
+for automated trading execution or real-time grid dispatch.
+
+## Citation & Attribution
+
+Data published under the ENTSO-E Transparency framework:
+- Transparency Platform: https://transparency.entsoe.eu/

config.json

@@ -0,0 +1,68 @@
+{
+  "target": "demand",
+  "country_code": "DE",
+  "zone_key": "DE",
+  "model_name": "lightgbm-quantile",
+  "model_version": "20260820180919",
+  "feature_names": [
+    "hour",
+    "quarter",
+    "day_of_week",
+    "day_of_year",
+    "month",
+    "is_weekend",
+    "is_holiday",
+    "is_morning_peak",
+    "is_evening_peak",
+    "sin_hour",
+    "cos_hour",
+    "sin_day_of_week",
+    "cos_day_of_week",
+    "sin_day_of_year",
+    "cos_day_of_year",
+    "lag_1h",
+    "lag_2h",
+    "lag_3h",
+    "lag_4h",
+    "lag_24h",
+    "lag_48h",
+    "lag_7d",
+    "lag_14d",
+    "diff_1h",
+    "diff_2h",
+    "diff_24h",
+    "diff_7d",
+    "acceleration_1h",
+    "ema_4step",
+    "ema_12step",
+    "rolling_std_4step",
+    "rolling_mean_24h",
+    "rolling_std_24h",
+    "rolling_min_24h",
+    "rolling_max_24h",
+    "rolling_mean_7d",
+    "rolling_std_7d"
+  ],
+  "n_estimators": 180,
+  "learning_rate": 0.04,
+  "num_leaves": 31,
+  "training_rows": 99380,
+  "start_year": 2023,
+  "end_year": 2026,
+  "metrics": {
+    "mae": 440.2182690041894,
+    "rmse": 663.2879914619127,
+    "wape": 0.007967997683390356,
+    "pinball_p10": 237.59465247807836,
+    "pinball_p90": 155.5080218314036,
+    "interval_coverage": 0.7652954435678635,
+    "winkler_score": 3931.0267430948206,
+    "mape": 0.778539019039999
+  },
+  "baseline_mae": 5542.1136924742395,
+  "calibrator": {
+    "is_fitted": true,
+    "q_correction": -34.60619320418482,
+    "target_coverage": 0.8
+  }
+}

inference.py

@@ -0,0 +1,29 @@
+# Standalone inference helper for Demand Forecaster
+import json
+from pathlib import Path
+import numpy as np
+import pandas as pd
+import joblib
+
+def load_forecaster(model_dir="."):
+    path = Path(model_dir)
+    config = json.loads((path / "config.json").read_text())
+    models = joblib.load(path / "models.joblib")
+    features = config["feature_names"]
+    q_correction = config["calibrator"]["q_correction"] if config.get("calibrator") else 0.0
+
+    def predict(df_features, apply_calibration=True):
+        X = df_features[features]
+        p10 = models[0.1].predict(X)
+        p50 = models[0.5].predict(X)
+        p90 = models[0.9].predict(X)
+        stacked = np.sort(np.vstack([p10, p50, p90]), axis=0)
+        p10, p50, p90 = stacked[0], stacked[1], stacked[2]
+        if apply_calibration:
+            p10 -= q_correction
+            p90 += q_correction
+            stacked_cal = np.sort(np.vstack([p10, p50, p90]), axis=0)
+            p10, p50, p90 = stacked_cal[0], stacked_cal[1], stacked_cal[2]
+        return pd.DataFrame({"p10": p10, "p50_point": p50, "p90": p90}, index=df_features.index)
+
+    return predict

models.joblib

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:c09e0e7b0e23ba4f6658a44807e35fa5872a868d4d3785256bcf8c6335e777f1
+size 1576238

sample_input.csv

@@ -0,0 +1,25 @@
+hour,quarter,day_of_week,day_of_year,month,is_weekend,is_holiday,is_morning_peak,is_evening_peak,sin_hour,cos_hour,sin_day_of_week,cos_day_of_week,sin_day_of_year,cos_day_of_year,lag_1h,lag_2h,lag_3h,lag_4h,lag_24h,lag_48h,lag_7d,lag_14d,diff_1h,diff_2h,diff_24h,diff_7d,acceleration_1h,ema_4step,ema_12step,rolling_std_4step,rolling_mean_24h,rolling_std_24h,rolling_min_24h,rolling_max_24h,rolling_mean_7d,rolling_std_7d
+6,0,5,319,11,1,0,0,0,1.0,6.123233995736766e-17,-0.9749279121818236,-0.2225209339563146,-0.7142921172691032,0.699847677146407,44845.90632,43470.27363,43228.13366,44723.16219,55164.24785,57809.17701,45797.761,42253.164,1375.6326899999985,1617.7726600000024,-2644.9291599999997,3544.5970000000016,1133.4927199999947,45013.65143129338,44766.49321885445,332.0792080029511,58120.48055354166,8661.062311121326,43109.34794,67379.83855,56797.70982903274,9024.021230559852
+6,1,5,319,11,1,0,0,0,1.0,6.123233995736766e-17,-0.9749279121818236,-0.2225209339563146,-0.7142921172691032,0.699847677146407,44675.1129,43956.67431,43109.34794,44358.08627,57437.36983,59688.19293,46130.684,42440.332,718.4385899999979,1565.7649600000004,-2250.823099999994,3690.351999999999,-128.88778000000457,45868.754974776035,45133.403537492224,1101.160538194595,58037.013495625,8728.20136809973,43109.34794,67379.83855,56799.724188095235,9021.71292869796
+6,2,5,319,11,1,0,0,0,1.0,6.123233995736766e-17,-0.9749279121818236,-0.2225209339563146,-0.7142921172691032,0.699847677146407,45003.10587,44459.59735,43403.51538,44139.36286,59143.01332,61682.27097,46724.045,43035.759,543.508520000003,1599.5904900000023,-2539.2576499999996,3688.286,-512.5734499999962,46536.18743286563,45503.2393194165,1246.4578525332379,57933.88814447916,8793.601217041185,43109.34794,67379.83855,56801.817420416664,9019.396681171937
+6,3,5,319,11,1,0,0,0,1.0,6.123233995736766e-17,-0.9749279121818236,-0.2225209339563146,-0.7142921172691032,0.699847677146407,45449.11325,44471.74481,43363.28636,44075.08203,60837.92453,63252.78388,47256.592,43616.517,977.3684400000056,2085.826890000004,-2414.859350000006,3640.074999999997,-131.09000999999262,47240.299887719375,45932.966896429345,1204.2390949354804,57820.90330333333,8847.418595774865,43109.34794,67379.83855,56804.15733644345,9016.981937571189
+7,0,5,319,11,1,0,0,0,0.9659258262890683,-0.25881904510252063,-0.9749279121818236,-0.2225209339563146,-0.7142921172691032,0.699847677146407,47151.41029,44845.90632,43470.27363,43228.13366,62386.79639,65055.89797,48312.398,43712.789,2305.503969999998,3681.1366599999965,-2669.101579999995,4599.609000000004,929.8712799999994,47942.656696631624,46404.23228313253,819.4241617229175,57697.551921875005,8887.371242302419,43109.34794,67379.83855,56806.74602678571,9014.48620273986
+7,1,5,319,11,1,0,0,0,0.9659258262890683,-0.25881904510252063,-0.9749279121818236,-0.2225209339563146,-0.7142921172691032,0.699847677146407,47537.33612,44675.1129,43956.67431,43109.34794,63383.74915,65738.53824,48758.569,43930.405,2862.22322,3580.661809999998,-2354.789089999991,4828.164000000004,2143.784630000002,48693.82827397897,46929.825107265984,975.1043226729313,57566.653893229166,8910.090444202244,43109.34794,67379.83855,56808.99035363096,9012.55576159426
+7,2,5,319,11,1,0,0,0,0.9659258262890683,-0.25881904510252063,-0.9749279121818236,-0.2225209339563146,-0.7142921172691032,0.699847677146407,48296.46857,45003.10587,44459.59735,43403.51538,63855.5222,66604.43633,49457.49,44623.031,3293.362699999998,3836.871220000001,-2748.9141299999974,4834.458999999995,2749.854179999995,49359.75961638738,47457.33764922506,907.864743791558,57430.97584614583,8919.742751760956,43109.34794,67379.83855,56811.37143641369,9010.636865550166
+7,3,5,319,11,1,0,0,0,0.9659258262890683,-0.25881904510252063,-0.9749279121818236,-0.2225209339563146,-0.7142921172691032,0.699847677146407,48996.19191,45449.11325,44471.74481,43363.28636,64384.72008,67014.77471,50594.347,45357.647,3547.078659999999,4524.447100000005,-2630.0546299999987,5236.700000000004,2569.7102199999936,49869.066409832434,47945.905180113514,721.1115564491922,57293.24151697916,8921.581255000769,43109.34794,67379.83855,56813.120746830355,9009.321079581974
+8,0,5,319,11,1,0,1,0,0.8660254037844387,-0.4999999999999998,-0.9749279121818236,-0.2225209339563146,-0.7142921172691032,0.699847677146407,49820.58564,47151.41029,44845.90632,43470.27363,65202.41408,67862.74414,51594.509,46134.116,2669.1753499999977,4974.679319999996,-2660.330059999993,5460.392999999996,363.6713799999998,50265.43730989947,48394.22648471143,447.96591020357806,57152.358950104164,8915.203076540178,43109.34794,67379.83855,56813.516054360116,9009.053631963381
+8,1,5,319,11,1,0,1,0,0.8660254037844387,-0.4999999999999998,-0.9749279121818236,-0.2225209339563146,-0.7142921172691032,0.699847677146407,50358.65663,47537.33612,44675.1129,43956.67431,65720.85739,68071.82898,52118.386,46695.388,2821.3205099999977,5683.543729999998,-2350.971590000001,5422.998,-40.902710000002116,51012.66678193969,48969.501023986595,785.362948912992,57016.224542916665,8890.732419058266,43109.34794,67379.83855,56814.31814065476,9008.612268516237
+8,2,5,319,11,1,0,1,0,0.8660254037844387,-0.4999999999999998,-0.9749279121818236,-0.2225209339563146,-0.7142921172691032,0.699847677146407,50633.0266,48296.46857,45003.10587,44459.59735,65716.09096,68254.32828,52947.774,47587.615,2336.55803,5629.920729999998,-2538.23732,5360.159,-956.8046699999977,51690.31420516382,49544.467841834805,998.5227481308469,56880.66129239584,8855.758211958908,43109.34794,67379.83855,56815.1937349107,9008.183750656593
+8,3,5,319,11,1,0,1,0,0.8660254037844387,-0.4999999999999998,-0.9749279121818236,-0.2225209339563146,-0.7142921172691032,0.699847677146407,50859.99366,48996.19191,45449.11325,44471.74481,65932.8991,67678.55822,53529.282,48442.32,1863.8017499999987,5410.880409999998,-1745.6591200000112,5086.9619999999995,-1683.2769100000005,52451.8622910983,50167.50116155253,1145.9338436436376,56754.39143260416,8814.777403946382,43109.34794,67379.83855,56816.155655178576,9007.804665602454
+9,0,5,319,11,1,0,1,0,0.7071067811865476,-0.7071067811865475,-0.9749279121818236,-0.2225209339563146,-0.7142921172691032,0.699847677146407,52133.51099,49820.58564,47151.41029,44845.90632,65754.34643,68150.21986,54274.782,49217.77,2312.925350000005,4982.100700000003,-2395.8734299999924,5057.012000000002,-356.2499999999927,53043.22277065899,50746.387673621364,820.9676682617471,56629.363978333335,8768.219288320763,43109.34794,67379.83855,56816.752353824406,9007.599889361314
+9,1,5,319,11,1,0,1,0,0.7071067811865476,-0.7071067811865475,-0.9749279121818236,-0.2225209339563146,-0.7142921172691032,0.699847677146407,52706.78534,50358.65663,47537.33612,44675.1129,65461.8974,67957.65274,54694.591,49552.093,2348.1287100000045,5169.449220000002,-2495.7553400000033,5142.498,-473.19179999999324,53791.665942395404,51387.60967767962,912.0378301316106,56516.4471478125,8719.132390173947,43109.34794,67379.83855,56817.7040631994,9007.364697096544
+9,2,5,319,11,1,0,1,0,0.7071067811865476,-0.7071067811865475,-0.9749279121818236,-0.2225209339563146,-0.7142921172691032,0.699847677146407,53594.18442,50633.0266,48296.46857,45003.10587,65572.78257,68087.16196,55261.132,50037.856,2961.1578200000004,5297.7158500000005,-2514.379390000002,5223.275999999998,624.5997900000002,54477.438109437244,52021.22301342122,880.9543458075615,56412.740886979176,8670.687379508972,43109.34794,67379.83855,56818.91166046131,9007.134027268838
+9,3,5,319,11,1,0,1,0,0.7071067811865476,-0.7071067811865475,-0.9749279121818236,-0.2225209339563146,-0.7142921172691032,0.699847677146407,53930.26349,50859.99366,48996.19191,45449.11325,66199.99543,69355.96608,55761.101,50531.286,3070.2698299999975,4934.071579999996,-3155.970650000003,5229.815000000002,1206.4680799999987,55081.895209662354,52631.58575904872,886.2514315553246,56312.9054525,8619.130845364418,43109.34794,67379.83855,56819.99417364583,9006.990240798747
+10,0,5,319,11,1,0,1,0,0.49999999999999994,-0.8660254037844387,-0.9749279121818236,-0.2225209339563146,-0.7142921172691032,0.699847677146407,54914.3307,52133.51099,49820.58564,47151.41029,66111.21083,69246.69143,55847.801,50798.68,2780.819709999996,5093.745060000001,-3135.48060000001,5049.120999999999,467.89435999999114,55536.55747779741,53183.426546887385,577.2390627431805,56208.93207177083,8558.59747507327,43109.34794,67379.83855,56820.67490263392,9006.927378978246
+10,1,5,319,11,1,0,1,0,0.49999999999999994,-0.8660254037844387,-0.9749279121818236,-0.2225209339563146,-0.7142921172691032,0.699847677146407,55506.09636,52706.78534,50358.65663,47537.33612,66159.33196,69456.69981,56066.312,50387.268,2799.311020000001,5147.4397300000055,-3297.3678500000096,5679.044000000002,451.1823099999965,56069.096022678445,53750.26920736625,565.8810776527902,56112.647623958335,8497.801729863984,43109.34794,67379.83855,56822.1929128125,9006.849131068186
+10,2,5,319,11,1,0,1,0,0.49999999999999994,-0.8660254037844387,-0.9749279121818236,-0.2225209339563146,-0.7142921172691032,0.699847677146407,55988.58086,53594.18442,50633.0266,48296.46857,66010.97078,69337.92876,56316.278,50328.922,2394.396440000004,5355.554260000004,-3326.957979999992,5987.356,-566.7613799999963,56476.654589607075,54263.765089309905,521.2826984545385,56018.154503958336,8435.113642260585,43109.34794,67379.83855,56823.71327061012,9006.807577904845
+10,3,5,319,11,1,0,1,0,0.49999999999999994,-0.8660254037844387,-0.9749279121818236,-0.2225209339563146,-0.7142921172691032,0.699847677146407,56218.55088,53930.26349,50859.99366,48996.19191,66093.00924,69273.47331,56513.973,50451.8,2288.287390000005,5358.557220000002,-3180.464070000002,6062.172999999995,-781.9824399999925,56863.316761764254,54752.92584787761,515.0570106516075,55928.90803770834,8373.371679988462,43109.34794,67379.83855,56825.39040159227,9006.817879710736
+11,0,5,319,11,1,0,1,0,0.258819045102521,-0.9659258262890682,-0.9749279121818236,-0.2225209339563146,-0.7142921172691032,0.699847677146407,56867.90384,54914.3307,52133.51099,49820.58564,66204.87532,69051.4665,56725.378,51795.587,1953.5731400000004,4734.3928499999965,-2846.5911799999885,4929.790999999997,-827.2465699999957,57218.33516105855,55214.14691128105,389.37814205827664,55842.01067854167,8309.825647628082,43109.34794,67379.83855,56827.23101135417,9006.88052798172
+11,1,5,319,11,1,0,1,0,0.258819045102521,-0.9659258262890682,-0.9749279121818236,-0.2225209339563146,-0.7142921172691032,0.699847677146407,57087.99244,55506.09636,52706.78534,50358.65663,66122.99543,69035.94704,57222.229,52210.049,1581.8960799999986,4381.2071,-2912.9516100000037,5012.18,-1217.4149400000024,57587.870924635135,55664.612704930114,448.42993572175476,55758.02421239583,8244.474290178683,43109.34794,67379.83855,56829.33933958333,9007.022471290016
+11,2,5,319,11,1,0,1,0,0.258819045102521,-0.9659258262890682,-0.9749279121818236,-0.2225209339563146,-0.7142921172691032,0.699847677146407,57443.31002,55988.58086,53594.18442,50633.0266,66189.35011,68814.8905,57619.989,53003.19,1454.7291599999953,3849.1255999999994,-2625.5403899999947,4616.798999999999,-939.6672800000088,57798.34243478109,56041.449165710095,331.5007239858158,55674.597694375,8178.74583091399,43109.34794,67379.83855,56830.666453720245,9007.146147318443
+11,3,5,319,11,1,0,1,0,0.258819045102521,-0.9659258262890682,-0.9749279121818236,-0.2225209339563146,-0.7142921172691032,0.699847677146407,57750.86276,56218.55088,53930.26349,50859.99366,66334.40707,68945.64319,57832.669,52968.243,1532.311880000001,3820.599270000006,-2611.2361200000014,4864.4259999999995,-755.9755100000039,58031.59096886866,56401.45141252392,260.1678840137753,55593.265545,8111.629723491022,43109.34794,67379.83855,56831.79960069941,9007.293493876698

sample_prediction.json

@@ -0,0 +1,80 @@
+{
+  "p10": [
+    45623.97,
+    47252.35,
+    47886.69,
+    48652.5,
+    49197.54,
+    50312.05,
+    51063.3,
+    51286.9,
+    51433.92,
+    52389.16,
+    52784.54,
+    53570.43,
+    53766.41,
+    54402.36,
+    55394.04,
+    55767.28,
+    55827.04,
+    56331.88,
+    56465.93,
+    56704.82,
+    56868.54,
+    57301.07,
+    57448.1,
+    57670.64
+  ],
+  "p50_point": [
+    47067.49,
+    47405.99,
+    47939.23,
+    48986.6,
+    50184.37,
+    50811.91,
+    51322.42,
+    51523.45,
+    51907.94,
+    52587.36,
+    53733.67,
+    54485.16,
+    54585.81,
+    55657.34,
+    56155.76,
+    56599.37,
+    56657.2,
+    57226.89,
+    57376.84,
+    57777.3,
+    57989.07,
+    58419.81,
+    58449.68,
+    58619.26
+  ],
+  "p90": [
+    47192.71,
+    48090.3,
+    48768.62,
+    49727.89,
+    50194.62,
+    50831.64,
+    51327.14,
+    51985.17,
+    52065.07,
+    53054.14,
+    53809.0,
+    54988.62,
+    54961.24,
+    55694.05,
+    56325.1,
+    56793.17,
+    56935.51,
+    57320.08,
+    57629.05,
+    58094.14,
+    58140.51,
+    58460.46,
+    58571.99,
+    58768.49
+  ]
+}