Time series model completed

BackEnd/time_series_models_builder.py

@@ -172,7 +172,104 @@ class TimeSeriesModel:
     #return self.feature_dataframes 
   
   
+ def plot_featured_data(self , feature='dayofweek' , district = "Monaragala"):
+    import pandas as pd
+    import matplotlib.pyplot as plt
+    import seaborn as sns
+    color_pal = sns.color_palette()
+
+    fig, ax = plt.subplots(figsize=(10, 8))
+    sns.boxplot(data=self.feature_dataframes[district], x=feature, y=district)
+    ax.set_title('Showing {} Data by {}'.format(district , feature))
+    plt.show()
+
+  def create_models(self):
+    import time
+    import xgboost as xgb
+
+    #This will create two dictionaries each contains feature datafranes for all the districts 
+    self.create_features(self.train , trainOrTest="train")
+    self.create_features(self.test , trainOrTest="test")
+
+    FEATURES = ['dayofyear', 'dayofweek', 'quarter', 'month', 'year']
+    
+    for district in self.df.columns:
+      TARGET = district
+
+      X_train = self.feature_train_dataframes[district][FEATURES]
+      y_train = self.feature_train_dataframes[district][TARGET]
+      
+      X_test = self.features_test_dataframes[district][FEATURES]
+      y_test = self.features_test_dataframes[district][TARGET]
+      self.model_x_tests[district] = X_test
+    
+
+      print("=========== Start Creating XGBooster for {} ===========".format(district))
+      reg = None
+      reg = xgb.XGBRegressor(base_score=0.5, booster='gbtree',    
+                       n_estimators=1000,
+                       early_stopping_rounds=50,
+                       objective='reg:linear',
+                       max_depth=3,
+                       learning_rate=0.01)
+      print("=========== Model Created ===========")
+      print("=========== Model Training : Started ===========")
+      reg.fit(X_train, y_train,
+        eval_set=[(X_train, y_train), (X_test, y_test)],
+        verbose=100)
+      print("=========== Model Training : Completed ===========")
+      self.ts_models[district] = reg
+      print("=========== Model Saved Successfully : {} District ===========".format(district))
+      time.sleep(3)
+    
+  def check_feature_importance(self , district="Monaragala"):
+      import pandas as pd
+      import matplotlib.pyplot as plt
+      import seaborn as sns
+      color_pal = sns.color_palette()
+
+      reg = self.ts_models[district]
+      fi = pd.DataFrame(data=reg.feature_importances_,
+             index=reg.feature_names_in_,
+             columns=['importance'])
+      fi.sort_values('importance').plot(kind='barh', title='Feature Importance')
+      plt.show()
+
+  def forecaset_data(self, district="Monaragala"):
+    import pandas as pd
+    import matplotlib.pyplot as plt
+    import seaborn as sns
+    color_pal = sns.color_palette()
+
+    reg = self.ts_models[district]
+    X_test = self.model_x_tests[district].copy()
+    test = self.features_test_dataframes[district].copy()
+    df = self.feature_dataframes[district].copy()
 
+    test['prediction'] = reg.predict(X_test) #get the predictions for testing dataset
+    
+    df = df.merge(test[['prediction']], how='left', left_index=True, right_index=True)
+    ax = df[[district]].plot(figsize=(15, 5))
+    df['prediction'].plot(ax=ax, style='.')
+    plt.legend(['Truth Data', 'Predictions'])
+    ax.set_title('Raw Dat and Prediction - {} District'.format(district))
+    plt.show()
+    #return test , df
+  
+  def forecast_week(self, starts_from , district="Monaragala"):
+    import pandas as pd
+    import matplotlib.pyplot as plt
+    import seaborn as sns
+    color_pal = sns.color_palette()
+  
+    end_date = pd.date_range(start=starts_from, periods=7+1, freq='D', closed='right')[-1]
+
+    ax = df.loc[(df.index > starts_from) & (df.index < end_date)][district] \
+    .plot(figsize=(15, 5), title='Week Of Data')
+    df.loc[(df.index > starts_from) & (df.index < end_date)]['prediction'] \
+    .plot(style='.')
+    plt.legend(['Truth Data','Prediction'])
+    plt.show()