Commit Live PR

q01_load_data/build.py

@@ -1,7 +1,16 @@
+# %load q01_load_data/build.py
 import pandas as pd
 import numpy as np
 from sklearn.model_selection import train_test_split
 
-"write your solution here"
-
+'write your solution here'
+def q01_load_data(path):
+    df=pd.read_csv(path)
+    #df['Month'] = pd.to_datetime(df['Month'])
+    #df.index = df['Month']
+    #df = df[['Sales']]
+    df.reset_index ( drop=True, inplace=True)
+    return df
+
+
 

q02_data_splitter/build.py

@@ -1,3 +1,4 @@
+# %load q02_data_splitter/build.py
 import pandas as pd
 import numpy as np
 import datetime
@@ -7,3 +8,11 @@
 from greyatomlib.time_series_101_project.q01_load_data.build import q01_load_data
 
 '''Write your solution here'''
+def q02_data_splitter(path):
+    tss=q01_load_data(path)
+    tss['Month']=pd.to_datetime(tss['Month'])
+    X_train = tss[tss['Month'] <datetime.datetime(1971, 10, 1, 0, 0, 0)]
+    X_valid = tss[tss['Month'] >=datetime.datetime(1971, 10, 1, 0, 0, 0)]
+    return X_train, X_valid
+
+

q03_stacked_point_plot/build.py

@@ -1,3 +1,4 @@
+# %load q03_stacked_point_plot/build.py
 import pandas as pd
 import numpy as np
 #import sys
@@ -8,5 +9,18 @@
 plt.switch_backend('agg')
 import seaborn as sns
 
-#"write your solution here"
-
+#'write your solution here'
+
+def q03_stacked_point_plot(path,x_column_name='month',y_column_name='Sales',hue='year',order_of_the_axis=['Jan', 'Feb', 'Mar', 'Apr', 'May', 'Jun', 'Jul', 'Aug','Sep', 'Oct', 'Nov', 'Dec']):
+    X_train,X_valid=q02_data_splitter(path)
+    X_train['year'] = X_train['Month'].dt.year
+    X_train['month'] = X_train['Month'].dt.strftime('%b')
+    plt.figure(figsize=(16, 7))
+    sns.pointplot(x=x_column_name, y=y_column_name, hue=hue, data=X_train, x_order=order_of_the_axis)
+    plt.xlabel('month')
+    plt.ylabel('Sales')
+    plt.title('Stacked point chart')
+    plt.legend(loc='upper right')
+
+
+

q04_boxplot/build.py

@@ -1,3 +1,4 @@
+# %load q04_boxplot/build.py
 import pandas as pd
 import numpy as np
 from sklearn.model_selection import train_test_split
@@ -9,6 +10,16 @@
 # sys.path.append('./')
 from greyatomlib.time_series_101_project.q02_data_splitter.build import q02_data_splitter
 
-"write your solution here"
+'write your solution here'
+def q04_boxplot(path,x='month',y='Sales',kind='box',order_of_the_axis=['Jan', 'Feb', 'Mar', 'Apr', 'May', 'Jun', 'Jul', 'Aug','Sep', 'Oct', 'Nov', 'Dec'],size=8):
+    X_train,X_valid=q02_data_splitter(path)
+    X_train['year']=X_train['Month'].dt.year
+    X_train[x]=X_train['Month'].dt.strftime('%b')
+
+    plt.figure(figsize=(16, 7))
+    sns.factorplot(x=x, y=y, data=X_train, kind=kind, order=order_of_the_axis, size=size, aspect=float(16/7))
+    plt.xlabel('month')
+    plt.ylabel('Sales')
+    plt.title('Boxplot using factorplot in seaborn')
 
 

q05_sarima_model/build.py

@@ -1,14 +1,29 @@
+# %load q05_sarima_model/build.py
 import pandas as pd
 import numpy as np
 from sklearn.model_selection import train_test_split
 import seaborn as sns
 import matplotlib.pyplot as plt
 # import sys
 # sys.path.append('./')
+path='data/perrin-freres-monthly-champagne.csv'
 from greyatomlib.time_series_101_project.q02_data_splitter.build import q02_data_splitter
 train, validation = q02_data_splitter(path)
+import datetime
 
 
-"write your solution here"
+'write your solution here'
+def q05_sarima_model(path):
+    train, validation = q02_data_splitter(path)
+    train.index=train['Month']
+    train=train[['Sales']]
+    train.index.name=None
+
+    validation.index=validation['Month']
+    validation=validation[['Sales']]
+    validation.index.name=None
+    return train, validation
+
 
 
+

q06_sarima_predictor/build.py

@@ -1,3 +1,4 @@
+# %load q06_sarima_predictor/build.py
 import pandas as pd
 import numpy as np
 from sklearn.model_selection import train_test_split
@@ -9,8 +10,24 @@
 from sklearn.metrics import mean_squared_error
 #import sys
 #sys.path.append('./')
+path='data/perrin-freres-monthly-champagne.csv'
 from greyatomlib.time_series_101_project.q05_sarima_model.build import q05_sarima_model
 tss, tss_valid = q05_sarima_model(path)
 
-    "write your solution here"
+'write your solution here'
+def q06_sarima_predictor(path):
+    tss, tss_valid = q05_sarima_model(path)
+    mod = sarimax.SARIMAX(tss.Sales, trend='n', order=(1,1,1), seasonal_order=(1,1,1,12))
+    results = mod.fit(disp=-1)
 
+    ## Forecasting 
+    pred = pd.DataFrame(results.forecast(len(tss_valid)))
+    pred.index = tss_valid.index
+
+
+    ## Measuring error. 
+    measure = math.pow(mean_squared_error(tss_valid.values, pred.values), 0.5)
+    return pred,measure
+
+
+