Create test_arima.py

Author: phumthep

scripts/test_arima.py

@@ -0,0 +1,70 @@
+import os
+from pownet.stochastic.solar import SolarTSModel
+from pownet.stochastic.demand import DemandTSModel
+import pandas as pd
+import matplotlib.pyplot as plt
+
+import logging
+
+# Show info
+logging.basicConfig(level=logging.INFO)
+
+# %% Demand
+
+data = pd.read_csv("../temp/hourly_demand_2023.csv")
+
+demand_model = DemandTSModel()
+demand_model.load_data(data)
+exog_vars = ["temp", "rhum", "prcp", "weekend"]
+
+# arima_order, seasonal_order = demand_model.find_best_model(
+#     target_column="demand",
+#     exog_vars=exog_vars,
+# )
+
+
+demand_model.fit(
+    target_column="demand",
+    exog_vars=exog_vars,
+    arima_order=(1, 0, 1),
+    seasonal_order=(0, 0, 0, 0),
+)
+predictions = demand_model.predict()
+
+data.index = pd.to_datetime(data["datetime"])
+exog_data = data[exog_vars].astype(float)
+synthetic = demand_model.get_synthetic(exog_data=exog_data)
+
+
+duration = 10
+# plt.plot(predictions[: 24 * 3], label="Predictions")
+plt.plot(synthetic[: 24 * duration], label="Synthetic")
+plt.plot(
+    data.set_index("datetime")["demand"].iloc[: 24 * duration],
+    label="Actual",
+)
+plt.legend()
+plt.show()
+
+"""# %% Solar
+data = pd.read_csv("../temp/merra_2019.csv")
+
+solar_model = SolarTSModel()
+solar_model.load_data(data)
+
+solar_model.fit(target_column="ground_irradiance", arima_order=(2, 1, 2))
+predictions = solar_model.predict()
+
+resids = solar_model.pred_residuals"""
+
+# synthetic = solar_model.get_synthetic()
+
+# duration = 30
+# # plt.plot(predictions[: 24 * 3], label="Predictions")
+# plt.plot(synthetic[: 24 * duration], label="Synthetic")
+# plt.plot(
+#     data.set_index("datetime")["ground_irradiance"].iloc[: 24 * duration],
+#     label="Actual",
+# )
+# plt.legend()
+# plt.show()