UP V1

mohammedmahdiali · web-flow · commit e49160a281db · 2023-06-27T20:40:10.000+03:00
diff --git a/rule_based_classification.py b/rule_based_classification.py
@@ -0,0 +1,210 @@
+#############################################
+# RULE BASED CLASSIFICATION - تصنيف العملاء
+#############################################
+
+# PRICE: صرف العميل
+# SOURCE: جهاز العميل
+# SEX: جنسية العميل
+# COUNTRY: بلد العميل
+# AGE: عمر العميل
+
+
+################# BEFORE - قبل #####################
+
+#    PRICE   SOURCE   SEX COUNTRY  AGE
+# 0     39  android  male     bra   17
+# 1     39  android  male     bra   17
+# 2     49  android  male     bra   17
+# 3     29  android  male     tur   17
+# 4     49  android  male     tur   17
+
+
+################# AFTER - بعد #####################
+
+#       customers_level_based        price segment
+# 0   BRA_ANDROID_FEMALE_0_18  1139.800000       A
+# 1  BRA_ANDROID_FEMALE_19_23  1070.600000       A
+# 2  BRA_ANDROID_FEMALE_24_30   508.142857       A
+# 3  BRA_ANDROID_FEMALE_31_40   233.166667       C
+# 4  BRA_ANDROID_FEMALE_41_66   236.666667       C
+
+import numpy as np
+import pandas as pd
+import seaborn as sns
+
+DATA_PATH = "../Data/persona.csv"
+df = pd.read_csv(DATA_PATH)
+
+
+################# DATA UNDERSTANDING #####################
+
+# Int64Index: 5000 entries, 0 to 4999
+# Data columns (total 5 columns):
+#  #   Column   Non-Null Count  Dtype 
+# ---  ------   --------------  ----- 
+#  0   PRICE    5000 non-null   int64 
+#  1   SOURCE   5000 non-null   object
+#  2   SEX      5000 non-null   object
+#  3   COUNTRY  5000 non-null   object
+#  4   AGE      5000 non-null   int64
+
+df.columns = ["price", "source", "sex", "country", "age"]
+
+df.price.value_counts()
+
+# 29    1305
+# 39    1260
+# 49    1031
+# 19     992
+# 59     212
+# 9      200
+
+df.describe()
+
+#              price          age
+# count  5000.000000  5000.000000
+# mean     34.132000    23.581400
+# std      12.464897     8.995908
+# min       9.000000    15.000000
+# 25%      29.000000    17.000000
+# 50%      39.000000    21.000000
+# 75%      39.000000    27.000000
+# max      59.000000    66.000000
+
+df.source.unique()
+
+# ['android' 'ios']
+
+df.sex.value_counts()
+
+# female    2621
+# male      2379
+
+df.country.value_counts()
+
+# usa    2065
+# bra    1496
+# deu     455
+# tur     451
+# fra     303
+# can     230
+
+df.groupby("country").agg({"price":"sum"}).sort_values(by="price", ascending=False)         
+
+# country  price     
+# usa      70225
+# bra      51354
+# tur      15689
+# deu      15485
+# fra      10177
+# can       7730
+
+df.groupby("country").agg({"price":"mean"}).sort_values(by="price", ascending=False)         
+             
+# country  price         
+# tur      34.787140
+# bra      34.327540
+# deu      34.032967
+# usa      34.007264
+# can      33.608696
+# fra      33.587459
+
+df.groupby("source").agg({"price":"mean"}).sort_values(by="price", ascending=False)             
+
+# source   price     
+# android  34.174849
+# ios      34.069102
+
+df.groupby(["country", "source"]).agg({"price":"mean"}).sort_values(by="price", ascending=False)                  
+
+# country source   price     
+# tur     android  36.229437
+# bra     android  34.387029
+# usa     ios      34.371703
+# fra     android  34.312500
+# deu     ios      34.268817
+# bra     ios      34.222222
+# can     ios      33.951456
+# deu     android  33.869888
+# usa     android  33.760357
+# can     android  33.330709
+# tur     ios      33.272727
+# fra     ios      32.776224
+
+
+################# DATA PROCESSING #####################
+
+agg_df = df.groupby([
+        "country", "source", "sex", "age"
+    ]).agg({
+            "price":"mean"
+        }).sort_values(by="price", ascending=False).reset_index()
+
+agg_df["new_age"] = pd.cut(agg_df["age"], bins=[0, 18, 23, 30, 40, 70], labels=['0_18', '19_23', '24_30', '31_40', '41_70'])
+
+#   country   source     sex  age  price new_age
+# 0     bra  android    male   46   59.0   41_70
+# 1     usa  android    male   36   59.0   31_40
+# 2     fra  android  female   24   59.0   24_30
+# 3     usa      ios    male   32   54.0   31_40
+# 4     deu  android  female   36   49.0   31_40
+
+agg_df["customer_level_based"] = ["_".join(col) for col in agg_df.drop(["age", "price"], axis=1).values]
+
+#   country   source     sex  age  price new_age      customer_level_based
+# 0     bra  android    male   46   59.0   41_70    bra_android_male_41_70
+# 1     usa  android    male   36   59.0   31_40    usa_android_male_31_40
+# 2     fra  android  female   24   59.0   24_30  fra_android_female_24_30
+# 3     usa      ios    male   32   54.0   31_40        usa_ios_male_31_40
+# 4     deu  android  female   36   49.0   31_40  deu_android_female_31_40
+
+agg_df = agg_df.groupby("customer_level_based").agg({"price":"mean"}).reset_index()
+
+#        customer_level_based  price
+# 0   bra_android_female_0_18  35.645303
+# 1  bra_android_female_19_23  34.077340
+# 2  bra_android_female_24_30  33.863946
+# 3  bra_android_female_31_40  34.898326
+# 4  bra_android_female_41_70  36.737179
+
+agg_df["segment"] = pd.qcut(agg_df["price"], 4, labels=["D", "C", "B", "A"])
+
+#    customer_level_based      price            segment
+# 0  bra_android_female_0_18   35.645303        B
+# 1  bra_android_female_19_23  34.077340        C
+# 2  bra_android_female_24_30  33.863946        C
+# 3  bra_android_female_31_40  34.898326        B
+# 4  bra_android_female_41_70  36.737179        A
+
+
+new_user = 'bra_android_female_31_40'
+agg_df[agg_df['customer_level_based'] == new_user]
+
+#        customer_level_based      price segment
+# 3  bra_android_female_31_40  34.898326       B
+
+
+############ Functionalization ############
+
+def rule_based_classifier(dataframe, bins=[0, 18, 23, 30, 40, 70], labels=['0_18', '19_23', '24_30', '31_40', '41_70'], export_to_csv=False):
+    
+    agg_df = df.groupby([
+            "country", "source", "sex", "age"
+        ]).agg({
+                "price":"mean"
+            }).sort_values(by="price", ascending=False).reset_index()
+
+    agg_df["new_age"] = pd.cut(agg_df["age"], bins=bins, labels=labels)
+    
+    agg_df["customer_level_based"] = ["_".join(col) for col in agg_df.drop(["age", "price"], axis=1).values]
+    
+    agg_df = agg_df.groupby("customer_level_based").agg({"price":"mean"}).reset_index()
+    
+    agg_df["segment"] = pd.qcut(agg_df["price"], 4, labels=["D", "C", "B", "A"])
+    
+    if export_to_csv:
+        agg_df.to_csv("rule_based_result.csv")
+    
+    return agg_df
+
+rule_based = rule_based_classifier(df, export_to_csv=True)
diff --git a/rule_based_result.csv b/rule_based_result.csv
@@ -0,0 +1,110 @@
+,customer_level_based,price,segment
+0,bra_android_female_0_18,35.645302694721295,B
+1,bra_android_female_19_23,34.077340379017514,C
+2,bra_android_female_24_30,33.86394557823129,C
+3,bra_android_female_31_40,34.89832643940475,B
+4,bra_android_female_41_70,36.73717948717949,A
+5,bra_android_male_0_18,34.80586076094243,B
+6,bra_android_male_19_23,31.673243243243245,D
+7,bra_android_male_24_30,33.413919413919416,C
+8,bra_android_male_31_40,34.327380952380956,B
+9,bra_android_male_41_70,40.041666666666664,A
+10,bra_ios_female_0_18,34.90041035353535,B
+11,bra_ios_female_19_23,36.40384615384615,A
+12,bra_ios_female_24_30,34.01587301587302,C
+13,bra_ios_female_31_40,29.875901875901878,D
+14,bra_ios_female_41_70,35.34090909090909,B
+15,bra_ios_male_0_18,34.66455912508544,B
+16,bra_ios_male_19_23,34.26604278074866,B
+17,bra_ios_male_24_30,33.86486486486486,C
+18,bra_ios_male_31_40,31.566137566137566,D
+19,bra_ios_male_41_70,31.08369408369408,D
+20,can_android_female_0_18,30.896551724137932,D
+21,can_android_female_19_23,32.01470588235294,D
+22,can_android_female_24_30,19.0,D
+23,can_android_female_41_70,34.0,C
+24,can_android_male_0_18,36.0,B
+25,can_android_male_19_23,40.111111111111114,A
+26,can_android_male_24_30,35.00961538461539,B
+27,can_android_male_41_70,37.57142857142857,A
+28,can_ios_female_0_18,31.352941176470587,D
+29,can_ios_female_24_30,31.727272727272727,D
+30,can_ios_male_0_18,36.74603174603175,A
+31,can_ios_male_24_30,31.17391304347826,D
+32,can_ios_male_31_40,29.0,D
+33,can_ios_male_41_70,31.0,D
+34,deu_android_female_0_18,30.912280701754383,D
+35,deu_android_female_19_23,33.333333333333336,C
+36,deu_android_female_24_30,33.56912878787879,C
+37,deu_android_female_31_40,37.333333333333336,A
+38,deu_android_female_41_70,36.083333333333336,A
+39,deu_android_male_0_18,37.2487922705314,A
+40,deu_android_male_19_23,36.07070707070707,A
+41,deu_android_male_24_30,19.0,D
+42,deu_android_male_31_40,35.972727272727276,B
+43,deu_ios_female_0_18,35.34522661523626,B
+44,deu_ios_female_19_23,34.0,C
+45,deu_ios_female_24_30,32.5,C
+46,deu_ios_female_31_40,33.0,C
+47,deu_ios_male_0_18,37.0,A
+48,deu_ios_male_19_23,39.0,A
+49,deu_ios_male_24_30,33.0,C
+50,deu_ios_male_31_40,31.352941176470587,D
+51,deu_ios_male_41_70,37.0,A
+52,fra_android_female_0_18,36.316958403914924,A
+53,fra_android_female_19_23,32.81944444444444,C
+54,fra_android_female_24_30,45.42857142857143,A
+55,fra_android_female_31_40,34.0,C
+56,fra_android_male_0_18,24.625,D
+57,fra_android_male_19_23,35.837606837606835,B
+58,fra_android_male_24_30,33.0,C
+59,fra_android_male_31_40,35.66666666666667,B
+60,fra_android_male_41_70,34.0,C
+61,fra_ios_female_0_18,31.312271062271062,D
+62,fra_ios_female_24_30,27.714285714285715,D
+63,fra_ios_female_31_40,32.81818181818181,C
+64,fra_ios_male_0_18,33.44444444444444,C
+65,fra_ios_male_19_23,39.888888888888886,A
+66,fra_ios_male_24_30,31.5,D
+67,fra_ios_male_31_40,33.6875,C
+68,fra_ios_male_41_70,39.0,A
+69,tur_android_female_0_18,37.685763888888886,A
+70,tur_android_female_19_23,33.44444444444445,C
+71,tur_android_female_24_30,30.785714285714285,D
+72,tur_android_female_31_40,41.83333333333333,A
+73,tur_android_female_41_70,34.11904761904762,B
+74,tur_android_male_0_18,39.87301587301587,A
+75,tur_android_male_19_23,27.333333333333332,D
+76,tur_android_male_24_30,39.43910256410256,A
+77,tur_android_male_31_40,29.0,D
+78,tur_android_male_41_70,39.0,A
+79,tur_ios_female_0_18,34.3921568627451,B
+80,tur_ios_female_19_23,36.06060606060606,A
+81,tur_ios_female_24_30,34.0,C
+82,tur_ios_female_31_40,32.333333333333336,D
+83,tur_ios_female_41_70,34.0,C
+84,tur_ios_male_0_18,35.55086848635236,B
+85,tur_ios_male_19_23,24.833333333333332,D
+86,tur_ios_male_24_30,45.0,A
+87,tur_ios_male_31_40,42.333333333333336,A
+88,tur_ios_male_41_70,25.05263157894737,D
+89,usa_android_female_0_18,33.649390332519474,C
+90,usa_android_female_19_23,33.627633891163306,C
+91,usa_android_female_24_30,31.269980506822616,D
+92,usa_android_female_31_40,32.8030303030303,C
+93,usa_android_female_41_70,39.25,A
+94,usa_android_male_0_18,34.577184614795236,B
+95,usa_android_male_19_23,34.27065165688917,B
+96,usa_android_male_24_30,31.886477667990274,D
+97,usa_android_male_31_40,35.904761904761905,B
+98,usa_android_male_41_70,34.33990414347557,B
+99,usa_ios_female_0_18,34.1824114487158,B
+100,usa_ios_female_19_23,35.37904761904762,B
+101,usa_ios_female_24_30,32.845238095238095,C
+102,usa_ios_female_31_40,28.51388888888889,D
+103,usa_ios_female_41_70,34.10372670807453,B
+104,usa_ios_male_0_18,33.983495499799844,C
+105,usa_ios_male_19_23,34.90187180941577,B
+106,usa_ios_male_24_30,34.83814333814334,B
+107,usa_ios_male_31_40,36.206324110671936,A
+108,usa_ios_male_41_70,35.75,B
