Updated test datasets by adding cyclical time of the day features. Re…

…named time of day features for easier readability in feature importance plots.

Signed-off-by: Egor Dmitriev <[email protected]>

openstef/feature_engineering/temporal_features.py

@@ -24,7 +24,7 @@ def add_time_of_the_day_cyclic_features(
     )
     period_of_the_day = 2 * np.pi * second_of_the_day / NUM_SECONDS_IN_A_DAY
 
-    data["sin_time_of_day"] = np.sin(period_of_the_day)
-    data["cos_time_of_day"] = np.cos(period_of_the_day)
+    data["time_of_day_sine"] = np.sin(period_of_the_day)
+    data["time_of_day_cosine"] = np.cos(period_of_the_day)
 
     return data

test/unit/data/input_data_with_holiday_features.csv

@@ -1,5 +1,5 @@
-,load,temp,humidity,pressure,T-1d,T-2d,T-3d,T-4d,T-5d,T-6d,T-7d,T-8d,T-9d,T-10d,T-11d,T-12d,T-13d,T-14d,IsWeekendDay,IsWeekDay,IsSunday,Month,Quarter,is_national_holiday,is_nieuwjaarsdag,is_goede_vrijdag,is_eerste_paasdag,is_tweede_paasdag,is_koningsdag,is_bevrijdingsdag,is_hemelvaart,is_eerste_pinksterdag,is_tweede_pinksterdag,is_eerste_kerstdag,is_tweede_kerstdag,is_bridgeday,is_schoolholiday,is_bouwvaknoord,is_voorjaarsvakantiemidden,is_voorjaarsvakantienoord,is_herfstvakantiezuid,is_kerstvakantie,is_zomervakantiemidden,is_herfstvakantiemidden,is_herfstvakantienoord,is_bouwvakmidden,is_zomervakantienoord,is_bouwvakzuid,is_zomervakantiezuid,is_meivakantie,is_voorjaarsvakantiezuid,saturation_pressure,vapour_pressure,dewpoint,air_density
-2020-02-01 10:00:00,10,9,1.0,,,,,,,,,,,,,,,,True,False,False,2,1,False,False,False,False,False,False,False,False,False,False,False,False,False,False,False,False,False,False,False,False,False,False,False,False,False,False,False,False,11.49201518556511,11.49201518556511,8.999999999999998,
-2020-02-01 10:10:00,15,9,0.02,,,,,,,,,,,,,,,,True,False,False,2,1,False,False,False,False,False,False,False,False,False,False,False,False,False,False,False,False,False,False,False,False,False,False,False,False,False,False,False,False,11.49201518556511,0.2298403037113022,-38.006190493385226,
-2022-12-26 10:00:00,20,9,0.03,,,,,,,,,,,,,,,,False,True,False,12,4,True,False,False,False,False,False,False,False,False,False,False,False,False,True,False,False,False,False,True,False,False,False,False,False,False,False,False,False,11.49201518556511,0.3447604555669533,-33.97264888737701,
-2020-04-27 11:00:00,15,9,0.04,,,,,,,,,,,,,,,,False,True,False,4,2,True,False,False,False,False,False,False,False,False,False,False,False,False,True,False,False,False,False,False,False,False,False,False,False,False,False,True,False,11.49201518556511,0.4596806074226044,-31.01205505051214,
+,load,temp,humidity,pressure,T-1d,T-2d,T-3d,T-4d,T-5d,T-6d,T-7d,T-8d,T-9d,T-10d,T-11d,T-12d,T-13d,T-14d,IsWeekendDay,IsWeekDay,IsSunday,Month,Quarter,is_national_holiday,is_nieuwjaarsdag,is_goede_vrijdag,is_eerste_paasdag,is_tweede_paasdag,is_koningsdag,is_bevrijdingsdag,is_hemelvaart,is_eerste_pinksterdag,is_tweede_pinksterdag,is_eerste_kerstdag,is_tweede_kerstdag,is_bridgeday,is_schoolholiday,is_bouwvaknoord,is_herfstvakantiemidden,is_meivakantie,is_bouwvakmidden,is_zomervakantienoord,is_zomervakantiemidden,is_voorjaarsvakantiezuid,is_bouwvakzuid,is_herfstvakantiezuid,is_voorjaarsvakantiemidden,is_herfstvakantienoord,is_kerstvakantie,is_zomervakantiezuid,is_voorjaarsvakantienoord,time_of_day_sine,time_of_day_cosine,saturation_pressure,vapour_pressure,dewpoint,air_density
+2020-02-01 10:00:00,10,9,1.0,,,,,,,,,,,,,,,,True,False,False,2,1,False,False,False,False,False,False,False,False,False,False,False,False,False,False,False,False,False,False,False,False,False,False,False,False,False,False,False,False,0.49999999999999994,-0.8660254037844387,11.49201518556511,11.49201518556511,8.999999999999998,
+2020-02-01 10:10:00,15,9,0.02,,,,,,,,,,,,,,,,True,False,False,2,1,False,False,False,False,False,False,False,False,False,False,False,False,False,False,False,False,False,False,False,False,False,False,False,False,False,False,False,False,0.4617486132350339,-0.8870108331782217,11.49201518556511,0.2298403037113022,-38.006190493385226,
+2022-12-26 10:00:00,20,9,0.03,,,,,,,,,,,,,,,,False,True,False,12,4,True,False,False,False,False,False,False,False,False,False,False,False,False,True,False,False,False,False,False,False,False,False,False,False,False,True,False,False,0.49999999999999994,-0.8660254037844387,11.49201518556511,0.3447604555669533,-33.97264888737701,
+2020-04-27 11:00:00,15,9,0.04,,,,,,,,,,,,,,,,False,True,False,4,2,True,False,False,False,False,False,False,False,False,False,False,False,False,True,False,False,True,False,False,False,False,False,False,False,False,False,False,False,0.258819045102521,-0.9659258262890682,11.49201518556511,0.4596806074226044,-31.01205505051214,

test/unit/feature_engineering/test_apply_features.py

@@ -93,11 +93,6 @@ def test_apply_features(self):
             pj={"model": "xgb", "lat": 52.132633, "lon": 5.291266},
         )
 
-        # Time of the day cyclic features are not in the expected output dataset
-        input_data_with_features = input_data_with_features.drop(
-            columns=time_of_the_day_cyclic_features_columns
-        )
-
         expected_output = TestData.load("input_data_with_features.csv")
 
         self.assertDataframeEqual(
@@ -131,11 +126,6 @@ def test_train_feature_applicator(self):
             pj={"model": "proleaf", "lat": 52.132633, "lon": 5.291266},
         )
 
-        # Time of the day cyclic features are not in the expected output dataset
-        input_data_with_features = input_data_with_features.drop(
-            columns=time_of_the_day_cyclic_features_columns
-        )
-
         expected_output = TestData.load("input_data_multi_horizon_features.csv")
 
         self.assertDataframeEqual(
@@ -205,11 +195,6 @@ def test_apply_holiday_features(self):
             data=input_data, horizon=24
         )
 
-        # Time of the day cyclic features are not in the expected output dataset
-        input_data_with_features = input_data_with_features.drop(
-            columns=time_of_the_day_cyclic_features_columns
-        )
-
         expected = TestData.load("../data/input_data_with_holiday_features.csv")
 
         self.assertDataframeEqual(

test/unit/feature_engineering/test_temporal.py

@@ -27,5 +27,5 @@ def test_add_time_of_the_day_cyclic_features(self):
             * np.pi
         )
 
-        assert np.allclose(output_data["sin_time_of_day"], np.sin(periods))
-        assert np.allclose(output_data["cos_time_of_day"], np.cos(periods))
+        assert np.allclose(output_data["time_of_day_sine"], np.sin(periods))
+        assert np.allclose(output_data["time_of_day_cosine"], np.cos(periods))