playing around with prophet

demo_prophet.py

@@ -0,0 +1,58 @@
+import pandas as pd
+import matplotlib.pyplot as plt
+from prophet import Prophet
+
+# df = pd.read_csv('https://raw.githubusercontent.com/facebook/prophet/main/examples/example_wp_log_peyton_manning.csv')
+
+cons_filename = 'pq_terheles_2021_adatok.tsv'
+
+df = pd.read_csv(cons_filename, sep='\t', skipinitialspace=True, na_values='n/a', decimal=',')
+df['Time'] = pd.to_datetime(df['Korrigált időpont'], format='%m/%d/%y %H:%M')
+df = df.set_index('Time')
+df['Consumption'] = df['Hatásos teljesítmény [kW]']
+
+df['ds'] = df.index
+df['y'] = df['Consumption']
+
+
+split_date = '2021-07-01'
+
+
+# TODO 15 minutes hardwired!
+forecast_horizon = 7 * 24 * 4
+
+# Split the data into training (past) and evaluation (future) sets
+train_data = df[df['ds'] <= split_date]
+eval_data = df[df['ds'] > split_date]
+
+
+# Initialize and train the Prophet model using the training data
+model = Prophet(seasonality_mode='multiplicative', growth='flat',
+    yearly_seasonality=True, weekly_seasonality=True, daily_seasonality=True)
+
+model.fit(train_data)
+
+# Create a DataFrame with future timestamps for the evaluation period
+future = model.make_future_dataframe(periods=forecast_horizon, freq='15T', include_history=False)
+
+# Make predictions for the evaluation period
+forecast = model.predict(future)
+
+# Calculate evaluation metrics (e.g., MAE, MSE, RMSE) by comparing eval_predictions with eval_data['y']
+
+# For example, you can calculate MAE as follows:
+from sklearn.metrics import mean_absolute_error
+
+eval_data = eval_data[eval_data['ds'] <= forecast['ds'].max()]
+
+mae = mean_absolute_error(eval_data['y'], forecast['yhat'])
+
+# Print or store the evaluation metrics
+print(f"Mean Absolute Error (MAE): {mae}")
+
+
+fig1 = model.plot(forecast)
+plt.show()
+
+fig2 = model.plot_components(forecast)
+plt.show()