customeranalytics 0.1__py3-none-any.whl

customeranalytics/exploratory_analysis/rfm.py

@@ -0,0 +1,202 @@
+import numpy as np
+import pandas as pd
+import sys, os, inspect
+import datetime
+
+currentdir = os.path.dirname(os.path.abspath(inspect.getfile(inspect.currentframe())))
+parentdir = os.path.dirname(currentdir)
+sys.path.insert(0, parentdir)
+
+from customeranalytics.configs import default_es_port, default_es_host, default_query_date
+from customeranalytics.utils import convert_to_date, current_date_to_day, convert_to_iso_format, get_index_group
+from customeranalytics.utils import calculate_time_diff, dimension_decision
+from customeranalytics.data_storage_configurations.query_es import QueryES
+
+
+class RFM:
+    """
+    RFM is a generally useful technique in order to classify customers according to their engagement with the business
+    Recency      : Last time attraction to the business.
+    Frequency(F) : how frequently users are engaged in the business.
+    Monetary(M)  : Average amount(value) per customer.
+    !!! The users, who have only 1 order will not be included in calculations !!!
+
+        !!!!
+        ******* ******** *****
+        Dimensional RFM:
+        RFM values must be created individually for dimensions. For instance, the Data set contains locations dimension.
+        In this case, each location of 'orders' and 'downloads' indexes must be created individually.
+        by using 'download_index' and 'order_index' dimensions can be assigned in order to create the RFM values.
+
+        download_index; downloads_location1 this will be the location dimension of
+                        parameters in order to query downloads indexes; 'location1'.
+        download_index; orders_location1 this will be the location dimension of
+                        parameters in order to query orders indexes; 'location1'.
+        ******* ******** *****
+        !!!
+    """
+
+    def __init__(self,
+                 host=None,
+                 port=None,
+                 download_index='downloads',
+                 order_index='orders'):
+        """
+
+        !!!!
+        ******* ******** *****
+        Dimensional RFM:
+        RFM values must be created individually for dimensions. For instance, the Data set contains locations dimension.
+        In this case, each location of 'orders' and 'downloads' indexes must be created individually.
+        by using 'download_index' and 'order_index' dimensions can be assigned in order to create the RFM values.
+
+        download_index; downloads_location1 this will be the location dimension of
+                        parameters in order to query downloads indexes; 'location1'.
+        download_index; orders_location1 this will be the location dimension of
+                        parameters in order to query orders indexes; 'location1'.
+        ******* ******** *****
+        !!!
+
+        :param download_index: elasticsearch port
+        :param order_index: elasticsearch port
+        :param host: elasticsearch host
+        :param port: elasticsearch port
+        """
+        self.port = default_es_port if port is None else port
+        self.host = default_es_host if host is None else host
+        self.download_index = download_index
+        self.order_index = order_index
+        self.query_es = QueryES(port=port, host=host)
+        self.orders_field_data = ["id", "session_start_date", "client", "payment_amount"]
+        self.orders = pd.DataFrame()
+        self.client_frequency = pd.DataFrame()
+        self.client_recency = pd.DataFrame()
+        self.client_monetary = pd.DataFrame()
+        self.rfm = pd.DataFrame()
+        self.max_order_date = datetime.datetime.now()
+
+    def dimensional_query(self, boolean_query=None):
+        if dimension_decision(self.order_index):
+            if boolean_query is None:
+                boolean_query = [{"term": {"dimension": self.order_index}}]
+            else:
+                boolean_query += [{"term": {"dimension": self.order_index}}]
+        return boolean_query
+
+    def get_data(self, start_date=None):
+        """
+        query orders index to collect the data with columns which are "session_start_date", "client", "payment_amount".
+        :param start_date:
+        :return:
+        """
+        start_date = default_query_date if start_date is None else start_date
+        if len(self.orders) == 0:
+            self.query_es = QueryES(port=self.port, host=self.host)
+            self.query_es.query_builder(fields=self.orders_field_data,
+                                        boolean_queries=self.dimensional_query([{"term": {"actions.purchased": True}}]),
+                                        date_queries=[{"range": {"session_start_date": {"gte": start_date}}}])
+            self.orders = pd.DataFrame(self.query_es.get_data_from_es())
+            self.orders['date'] = self.orders['session_start_date'].apply(lambda x: convert_to_date(x))
+
+    def frequency(self):
+        """
+        Frequency of users;
+            -   assign dates of next orders per user as a column.
+                So, each row will have a current order date and the next order date per user.
+            -   Calculate the hour difference from the current order date to the next order date.
+            -   Calculate the average hourly difference per user.
+        User has only 1 order will not be included in calculations.
+        """
+        self.orders['next_order_date'] = self.orders.sort_values(
+            by=['client', 'date'], ascending=True).groupby(['client'])['date'].shift(-1)
+        self.orders['diff_hours'] = self.orders.apply(
+            lambda row: calculate_time_diff(row['date'], row['next_order_date'], 'hour'), axis=1)
+        self.client_frequency = self.orders.query("next_order_date == next_order_date").groupby("client").agg(
+            {"diff_hours": "mean"}).reset_index().rename(columns={"diff_hours": "frequency"})
+
+    def recency(self):
+        """
+        The recency of users;
+            -   Calculate the last transaction (purchased) date of the whole population.
+            -   Find each user of maximum transaction (purchased) date.
+            -   Calculate the hour difference from the maximum transaction date to each user of the maximum transaction date.
+        """
+        self.max_order_date = max(self.orders['date'])
+        self.client_recency = self.orders.groupby("client").agg({"date": "max"}).reset_index()
+        self.client_recency['recency'] = self.client_recency.apply(
+            lambda row: calculate_time_diff(row['date'], self.max_order_date, 'hour'), axis=1)
+        self.client_recency = self.client_recency.drop('date', axis=1)
+
+    def monetary(self):
+        """
+        Monetary of users;
+            -   Calculate the average purchased amount per user
+        """
+        self.client_monetary = self.orders.groupby("client").agg({"payment_amount": "mean"}).reset_index().rename(
+            columns={"payment_amount": "monetary"})
+
+    def execute_rfm(self, start_date):
+        """
+        1.  Execute R, F, M calculations.
+        2.  Merge data-frames (R, F, M data-frames).
+        3.  Insert into the reports index with report_name 'rfm'.
+        """
+        self.get_data(start_date=start_date)
+        self.frequency()
+        self.recency()
+        self.monetary()
+        self.rfm = pd.merge(self.client_frequency, self.client_recency, on='client', how='left')
+        self.rfm = pd.merge(self.rfm, self.client_monetary, on='client', how='left')
+        self.insert_into_reports_index(self.rfm, start_date, index=self.order_index)
+
+    def insert_into_reports_index(self, rfm, start_date, index='orders'):
+        """
+        via query_es.py, each report can be inserted into the reports index with the given format.
+        {"id": unique report id,
+         "report_date": start_date or current date,
+         "report_name": "rfm",
+         "index": "main",
+         "report_types": {},
+         "data": rfm.fillna(0.0).to_dict("records") -  dataframe to list of dictionary
+         }
+         !!! null values are assigned to 0.
+
+        :param rfm: data set, data frame
+        :param start_date: data start date
+        :param index: dimensionality of data index orders_location1 ;  dimension = location1
+        """
+
+        list_of_obj = [{"id": np.random.randint(200000000),
+                        "report_date": current_date_to_day().isoformat() if start_date is None else start_date,
+                        "report_name": "rfm",
+                        "index": get_index_group(index),
+                        "report_types": {},
+                        "data": rfm.fillna(0).to_dict("records")}]
+        self.query_es.insert_data_to_index(list_of_obj, index='reports')
+
+    def fetch(self, start_date=None):
+        """
+        Collect RFM values for each user. Collecting stored RFM is useful in order to initialize Customer Segmentation.
+        :return: data-frame
+        """
+
+        boolean_queries, date_queries = [], []
+        boolean_queries = [{"term": {"report_name": "rfm"}}, {"term": {"index": get_index_group(self.order_index)}}]
+
+        if start_date is not None:
+            date_queries = [{"range": {"report_date": {"gte": convert_to_iso_format(start_date)}}}]
+
+        self.query_es = QueryES(port=self.port,
+                                host=self.host)
+        self.query_es.query_builder(fields=None, _source=True,
+                                    date_queries=date_queries,
+                                    boolean_queries=boolean_queries)
+        _res = self.query_es.get_data_from_es(index="reports")
+        _data = pd.DataFrame()
+        if len(_res) != 0:
+            _data = pd.DataFrame(_res[0]['_source']['data'])
+        return _data
+
+
+
+

customeranalytics/exploratory_analysis/sample_data/sample_data_avg_order_count_per_promo_per_cust.csv

@@ -0,0 +1,99 @@
+promotion_id,average_order_count
+promo_0,1.0411764705882354
+promo_1,1.0467479674796747
+promo_10,1.056969696969697
+promo_11,1.0535917901938427
+promo_12,1.0515826494724503
+promo_13,1.036
+promo_14,1.0472589792060492
+promo_15,1.0474452554744527
+promo_16,1.0476190476190477
+promo_17,1.0305343511450382
+promo_18,1.0395569620253164
+promo_19,1.0625
+promo_2,1.0574018126888218
+promo_20,1.0476923076923077
+promo_21,1.0354838709677419
+promo_22,1.0432632880098887
+promo_23,1.0373831775700935
+promo_24,1.041044776119403
+promo_25,1.048854961832061
+promo_26,1.0447284345047922
+promo_27,1.0296846011131726
+promo_28,1.0359550561797752
+promo_29,1.035031847133758
+promo_3,1.0568047337278106
+promo_30,1.0486425339366516
+promo_31,1.0683453237410072
+promo_32,1.0454545454545454
+promo_33,1.062043795620438
+promo_34,1.0361842105263157
+promo_35,1.0470809792843692
+promo_36,1.0493601462522852
+promo_37,1.0358490566037737
+promo_38,1.0465380249716232
+promo_39,1.0421179302045729
+promo_4,1.0407407407407407
+promo_40,1.0455373406193078
+promo_41,1.0520094562647755
+promo_42,1.0400801603206413
+promo_43,1.0488721804511278
+promo_44,1.0424028268551238
+promo_45,1.0629139072847682
+promo_46,1.0413016270337923
+promo_47,1.038327526132404
+promo_48,1.0483619344773791
+promo_49,1.0502901353965184
+promo_5,1.0495356037151702
+promo_50,1.0671140939597314
+promo_51,1.057057057057057
+promo_52,1.0461847389558232
+promo_53,1.048689138576779
+promo_54,1.038181818181818
+promo_55,1.0615735461801596
+promo_56,1.044871794871795
+promo_57,1.0557768924302788
+promo_58,1.0372670807453417
+promo_59,1.04
+promo_6,1.0363288718929253
+promo_60,1.0388349514563107
+promo_61,1.0595238095238095
+promo_63,1.034013605442177
+promo_64,1.0441988950276244
+promo_65,1.0666666666666667
+promo_66,1.0303030303030303
+promo_67,1.0490367775831875
+promo_68,1.0618932038834952
+promo_69,1.0452830188679245
+promo_7,1.025735294117647
+promo_70,1.0141843971631206
+promo_71,1.0551558752997603
+promo_72,1.0520833333333333
+promo_73,1.0506634499396863
+promo_74,1.0621019108280254
+promo_75,1.0323741007194245
+promo_76,1.0429378531073445
+promo_77,1.03954802259887
+promo_78,1.0688018979833926
+promo_79,1.0381355932203389
+promo_8,1.0390879478827362
+promo_80,1.0526315789473684
+promo_81,1.044280442804428
+promo_82,1.041522491349481
+promo_83,1.0603948896631823
+promo_84,1.0264550264550265
+promo_85,1.0414691943127963
+promo_86,1.0341614906832297
+promo_87,1.0456140350877192
+promo_88,1.059610705596107
+promo_9,1.06006006006006
+promo_90,1.0430733410942956
+promo_91,1.0526315789473684
+promo_92,1.048602673147023
+promo_93,1.046218487394958
+promo_94,1.053763440860215
+promo_95,1.0588235294117647
+promo_96,1.0595238095238095
+promo_97,1.055045871559633
+promo_98,1.0562874251497005
+promo_99,1.037037037037037

customeranalytics/exploratory_analysis/sample_data/sample_data_chart_1_search.csv

@@ -0,0 +1,2 @@
+products,average_product_sold_per_user_kpi,total_product_revenue_kpi,total_product_discount_kpi,total_product_cust_kpi
+p_127,1.2003129890453834,113401.1849999997,5743.149899999994,3195

customeranalytics/exploratory_analysis/sample_data/sample_data_chart_2_search.csv

@@ -0,0 +1,46 @@
+daily,products,payment_amount,order_count
+2021-06-01,p_127,2443.0200000000004,80
+2021-06-02,p_127,2977.304999999999,100
+2021-06-03,p_127,2321.125,83
+2021-06-04,p_127,2604.144999999999,91
+2021-06-05,p_127,2648.175,88
+2021-06-06,p_127,2623.79,89
+2021-06-07,p_127,2113.52,73
+2021-06-08,p_127,2761.62,87
+2021-06-09,p_127,2822.115000000001,97
+2021-06-10,p_127,2502.4900000000007,84
+2021-06-11,p_127,2129.9949999999994,77
+2021-06-12,p_127,2160.69,71
+2021-06-13,p_127,2481.3749999999995,84
+2021-06-14,p_127,2270.1150000000007,82
+2021-06-15,p_127,2377.2850000000008,84
+2021-06-16,p_127,2652.715,90
+2021-06-17,p_127,2814.22,95
+2021-06-18,p_127,3089.0200000000013,104
+2021-06-19,p_127,2539.245000000001,84
+2021-06-20,p_127,2605.994999999999,83
+2021-06-21,p_127,2836.4400000000014,95
+2021-06-22,p_127,2874.02,95
+2021-06-23,p_127,2083.7850000000008,71
+2021-06-24,p_127,2325.600000000001,79
+2021-06-25,p_127,2097.5799999999995,74
+2021-06-26,p_127,2083.7200000000007,71
+2021-06-27,p_127,2935.070000000001,93
+2021-06-28,p_127,2390.0,81
+2021-06-29,p_127,2836.7900000000004,104
+2021-06-30,p_127,2458.58,75
+2021-07-01,p_127,2536.115,82
+2021-07-02,p_127,2478.3099999999995,87
+2021-07-03,p_127,2862.2250000000013,92
+2021-07-04,p_127,2605.034999999999,86
+2021-07-05,p_127,2308.155,82
+2021-07-06,p_127,1989.8450000000007,66
+2021-07-07,p_127,2556.01,85
+2021-07-08,p_127,2267.88,70
+2021-07-09,p_127,2184.734999999999,84
+2021-07-10,p_127,2292.044999999999,77
+2021-07-11,p_127,3018.315,97
+2021-07-12,p_127,2721.760000000001,95
+2021-07-13,p_127,2337.795,86
+2021-07-14,p_127,2592.885000000001,96
+2021-07-15,p_127,2790.530000000001,86

customeranalytics/exploratory_analysis/sample_data/sample_data_chart_3_search.csv

@@ -0,0 +1,2 @@
+products,accept_Ratio,mean_control,mean_validation,is_orders_increased_per_products,diff
+p_127,0.0,16.018044714064835,28.024737393941244,True,12.006692679876409

customeranalytics/exploratory_analysis/sample_data/sample_data_chart_4_search.csv

@@ -0,0 +1,2 @@
+products,accept_Ratio,mean_control,mean_validation,is_orders_increased_per_products,diff
+p_127,0.0,16.018044714064835,28.024737393941244,True,12.006692679876409

customeranalytics/exploratory_analysis/sample_data/sample_data_churn.csv

@@ -0,0 +1,2 @@
+churn
+0.7976236730294409

customeranalytics/exploratory_analysis/sample_data/sample_data_churn_weekly.csv

@@ -0,0 +1,10 @@
+weekly,churn
+2021-04-26T00:00:00,0.0
+2021-05-03T00:00:00,0.22315777576853527
+2021-05-10T00:00:00,0.5375455650060753
+2021-05-17T00:00:00,0.6788188716595183
+2021-05-24T00:00:00,0.7658633650232147
+2021-05-31T00:00:00,0.8036602700781805
+2021-06-07T00:00:00,0.8307974567424794
+2021-06-14T00:00:00,0.8633264453405916
+2021-06-21T00:00:00,0.925724927738967