sesd_algorithm.py

def sesd_anomaly_detection(dataframe, 
                           column_name, 
                           desired_frequency, 
                           max_anomalies,
                           alpha_level):
    """
    In this definition, time series anomalies are detected using the S-ESD algorithm.
    Arguments:
        dataframe: Pandas dataframe
        column_name: string. Name of the column that we want to detect anomalies in
        desired_frequency: Integer. Time frequency of the series. If we want to detect 
        a yearly trend, we'd set the value equal to 365.
        max_anomalies: Integer. Max number of anomalies to look for in the time series
        sequence.
        alpha_level: The significance level.
    Outputs:
        df: Pandas dataframe with column for detected S-ESD anomalies (True/False)
    """
    series=np.array(dataframe[column_name])
    #Implement SESD algorithm on the time series
    outliers_indices = sesd.seasonal_esd(series, 
                                         hybrid=False, 
                                         seasonality=desired_frequency,
                                         max_anomalies=max_anomalies, 
                                         alpha=alpha_level)
    #Create a column for SESD anomalies
    sesd_anomaly_column=column_name+'_SESD_Anomaly'
    #Create a detected anomaly column, and mark as False if normal, and True if anomalous
    dataframe[sesd_anomaly_column]=False
    dataframe.loc[dataframe.index.isin(outliers_indices), sesd_anomaly_column]=True
    return dataframe

##EXECUTE IN MAIN BLOCK

#Implement the SESD algorithm on the time series
gasoline_price_df=sesd_anomaly_detection(dataframe=gasoline_price_df, 
                           column_name='Gasoline_Price', 
                           desired_frequency=365, 
                           max_anomalies=20,
                           alpha_level=10)
#Re-plot time series with color coding for anomaly column
scatterplot_with_color_coding(gasoline_price_df['Date'],
                                  gasoline_price_df['Gasoline_Price'], 
                                  gasoline_price_df['Gasoline_Price_SESD_Anomaly'],
                                  'Date', 
                                  'Gasoline Price (Dollars Per Gallon)', 
                                  'Gasoline Prices, Color-Coded on SESD Anomalies')