Update eda
Browse files
app.py
CHANGED
@@ -2,11 +2,11 @@
|
|
2 |
import streamlit as st
|
3 |
|
4 |
# Import the created Streamlit Page
|
5 |
-
|
6 |
import prediction
|
7 |
|
8 |
# Navigation button
|
9 |
-
navigasi = st.sidebar.selectbox(label='Select Page:', options=['Home Page', '
|
10 |
|
11 |
# Looping for navigation
|
12 |
if navigasi == 'Home Page':
|
@@ -18,9 +18,9 @@ if navigasi == 'Home Page':
|
|
18 |
st.write('**Classification Model for Predicting Customer Churn**')
|
19 |
st.caption('Please select another menu in the Select Box on the left of your screen to start!')
|
20 |
|
21 |
-
#
|
22 |
-
|
23 |
-
|
24 |
|
25 |
# Displays the Predict page
|
26 |
elif navigasi == 'Prediction':
|
|
|
2 |
import streamlit as st
|
3 |
|
4 |
# Import the created Streamlit Page
|
5 |
+
import eda
|
6 |
import prediction
|
7 |
|
8 |
# Navigation button
|
9 |
+
navigasi = st.sidebar.selectbox(label='Select Page:', options=['Home Page', 'Exploratory Data Analysis', 'Prediction'])
|
10 |
|
11 |
# Looping for navigation
|
12 |
if navigasi == 'Home Page':
|
|
|
18 |
st.write('**Classification Model for Predicting Customer Churn**')
|
19 |
st.caption('Please select another menu in the Select Box on the left of your screen to start!')
|
20 |
|
21 |
+
# Displays the EDA page
|
22 |
+
elif navigasi == 'Exploratory Data Analysis':
|
23 |
+
eda.run()
|
24 |
|
25 |
# Displays the Predict page
|
26 |
elif navigasi == 'Prediction':
|
eda.py
ADDED
@@ -0,0 +1,152 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
|
2 |
+
import streamlit as st
|
3 |
+
import pandas as pd
|
4 |
+
import matplotlib.pyplot as plt
|
5 |
+
import seaborn as sns
|
6 |
+
|
7 |
+
# Function to run the EDA
|
8 |
+
def run():
|
9 |
+
# Load dataset
|
10 |
+
df = pd.read_csv('Churn_Modelling.csv')
|
11 |
+
|
12 |
+
# Set the title in the Streamlit app
|
13 |
+
st.title('Exploratory Data Analysis of Customer Churn')
|
14 |
+
st.write('---')
|
15 |
+
|
16 |
+
# 1. Distribution of Customer Churn
|
17 |
+
st.subheader('1. Distribution of Customer Churn')
|
18 |
+
value_counts = df['Exited'].value_counts()
|
19 |
+
fig, (ax1, ax2) = plt.subplots(1, 2, figsize=(14, 6))
|
20 |
+
colors_pie = ['pink', 'grey']
|
21 |
+
ax1.pie(value_counts, labels=['Not Churn', 'Churn'], autopct='%1.1f%%', colors=colors_pie)
|
22 |
+
ax1.set_title('Distribution of Customer Churn', fontsize=12)
|
23 |
+
colors_bar = ['pink', 'grey']
|
24 |
+
value_counts.index = ['Not Churn', 'Churn']
|
25 |
+
value_counts.plot(kind='bar', ax=ax2, color=colors_bar)
|
26 |
+
ax2.set_title('Count of Customer Churn', fontsize=12)
|
27 |
+
ax2.set_xlabel('Churn Status', fontsize=10)
|
28 |
+
ax2.set_ylabel('Count', fontsize=10)
|
29 |
+
ax2.tick_params(axis='x', rotation=0)
|
30 |
+
for i, v in enumerate(value_counts):
|
31 |
+
ax2.text(i, v + 10, str(v), ha='center', va='bottom')
|
32 |
+
st.pyplot(fig)
|
33 |
+
st.write('---')
|
34 |
+
|
35 |
+
# 2. Customer Churn by Gender
|
36 |
+
st.subheader('2. Customer Churn by Gender')
|
37 |
+
gender_group = df.groupby(['Exited', 'Gender']).size().unstack()
|
38 |
+
fig, ax = plt.subplots(figsize=(8, 5))
|
39 |
+
bar_chart = gender_group.plot(kind='bar', color=['pink', 'grey'], ax=ax)
|
40 |
+
for container in bar_chart.containers:
|
41 |
+
bar_chart.bar_label(container)
|
42 |
+
bar_chart.set_xticklabels(['Not Churn', 'Churn'], rotation=0)
|
43 |
+
bar_chart.set_xlabel('Churn Status')
|
44 |
+
bar_chart.set_ylabel('Count')
|
45 |
+
bar_chart.set_title('Customer Churn by Gender')
|
46 |
+
st.pyplot(fig)
|
47 |
+
st.write('**Churn by Gender Statistics**')
|
48 |
+
st.write('Percentage of Male customers who churned: 16.5%')
|
49 |
+
st.write('Percentage of Female customers who churned: 25.1%')
|
50 |
+
st.write('---')
|
51 |
+
|
52 |
+
|
53 |
+
# 3 Customer Churn by Age
|
54 |
+
st.subheader('3. Customer Churn by Age')
|
55 |
+
churn_customers = df[df['Exited'] == 1]
|
56 |
+
not_churn_customers = df[df['Exited'] == 0]
|
57 |
+
plt.figure(figsize=(10, 6))
|
58 |
+
plt.hist(not_churn_customers['Age'], bins=30, alpha=0.5, label='Not Churn', color='pink')
|
59 |
+
plt.hist(churn_customers['Age'], bins=30, alpha=0.5, label='Churn', color='grey')
|
60 |
+
plt.title('Customer Churn by Age', fontsize=14)
|
61 |
+
plt.xlabel('Age')
|
62 |
+
plt.ylabel('Frequency')
|
63 |
+
plt.legend(loc='upper right')
|
64 |
+
st.pyplot()
|
65 |
+
st.write('---')
|
66 |
+
|
67 |
+
# 4. Customer Churn by Geography
|
68 |
+
st.subheader('4. Customer Churn by Geography')
|
69 |
+
geo_group = df.groupby(['Exited', 'Geography']).size().unstack()
|
70 |
+
fig, ax = plt.subplots(figsize=(8, 5))
|
71 |
+
bar_char = geo_group.plot(kind='bar', color=['pink', 'grey', 'lightblue'], ax=ax)
|
72 |
+
for container in bar_char.containers:
|
73 |
+
bar_char.bar_label(container)
|
74 |
+
bar_char.set_xticklabels(['Not Churn', 'Churn'], rotation=0)
|
75 |
+
bar_char.set_xlabel('Churn Status', fontsize=9)
|
76 |
+
bar_char.set_ylabel('Count', fontsize=9)
|
77 |
+
bar_char.set_title('Customer Churn by Geography', fontsize=10)
|
78 |
+
st.pyplot()
|
79 |
+
st.write('**Churn by Geography Statistics**')
|
80 |
+
st.write('Percentage of customers who churned in France: 16.2 %')
|
81 |
+
st.write('Percentage of customers who churned in Germany: 32.5 %')
|
82 |
+
st.write('Percentage of customers who churned in Spain: 16.7 %')
|
83 |
+
st.write('---')
|
84 |
+
|
85 |
+
# 5. Customer Churn by Tenure
|
86 |
+
st.subheader('5. Customer Churn by Tenure')
|
87 |
+
churn_by_tenure = df.groupby(['Tenure', 'Exited']).size().unstack()
|
88 |
+
fig, ax = plt.subplots(figsize=(8, 5))
|
89 |
+
churn_by_tenure.plot(kind='bar', stacked=True, color=['pink', 'grey'], ax=ax)
|
90 |
+
plt.title('Customer Churn by Tenure', fontsize=14)
|
91 |
+
plt.xticks(rotation=0)
|
92 |
+
plt.xlabel('Tenure (years)')
|
93 |
+
plt.ylabel('Number of Customers')
|
94 |
+
plt.legend(['Not Churn', 'Churn'])
|
95 |
+
for container in ax.containers:
|
96 |
+
ax.bar_label(container, label_type='center')
|
97 |
+
st.pyplot(fig)
|
98 |
+
st.write('---')
|
99 |
+
|
100 |
+
# 6. Customer Churn by Number of Products
|
101 |
+
st.subheader('6. Customer Churn by Number of Products')
|
102 |
+
plt.figure(figsize=(8, 5))
|
103 |
+
ax = sns.countplot(data=df, x='NumOfProducts', hue='Exited', palette=['grey', 'pink'])
|
104 |
+
plt.title('Customer Churn by Number of Products', fontsize=14)
|
105 |
+
plt.xlabel('Number of Products', fontsize=10)
|
106 |
+
plt.ylabel('Number of Customers', fontsize=10)
|
107 |
+
for container in ax.containers:
|
108 |
+
ax.bar_label(container)
|
109 |
+
st.pyplot()
|
110 |
+
st.write('**Churn by Number of Products Statistics**')
|
111 |
+
st.write('Percentage of customers who churned with 1 products: 27.7 %')
|
112 |
+
st.write('Percentage of customers who churned with 2 products: 7.6 %')
|
113 |
+
st.write('Percentage of customers who churned with 3 products: 82.7 %')
|
114 |
+
st.write('Percentage of customers who churned with 4 products: 100 %')
|
115 |
+
st.write('---')
|
116 |
+
|
117 |
+
# 7. Customer Churn by Has Credit Card or Not
|
118 |
+
st.subheader('7. Customer Churn by Has Credit Card or Not')
|
119 |
+
churn_by_HasCrCard = df.groupby(['HasCrCard', 'Exited']).size().unstack()
|
120 |
+
fig, ax = plt.subplots(figsize=(8, 4))
|
121 |
+
churn_by_HasCrCard.plot(kind='bar', stacked=True, color=['pink', 'grey'], ax=ax)
|
122 |
+
plt.title('Customer Churn by HasCrCard', fontsize=11)
|
123 |
+
plt.xticks(rotation=0)
|
124 |
+
plt.xlabel('HasCrCard', fontsize=10)
|
125 |
+
plt.ylabel('Number of Customers', fontsize=10)
|
126 |
+
plt.legend(['Not Churn', 'Churn'])
|
127 |
+
ax.set_xticklabels(['Has Credit Card', 'No Credit Card'])
|
128 |
+
for container in ax.containers:
|
129 |
+
ax.bar_label(container, label_type='center')
|
130 |
+
st.pyplot(fig)
|
131 |
+
st.write('**Churn by Has Credit Card or Not Statistics**')
|
132 |
+
st.write('Percentage of customers who churned with No Credit Card: 20.8 %')
|
133 |
+
st.write('Percentage of customers who churned with Has Credit Card: 20.2 %')
|
134 |
+
st.write('---')
|
135 |
+
|
136 |
+
# 8. Customer Churn by Is Active Member or Not
|
137 |
+
st.subheader('8. Customer Churn by Is Active Member or Not')
|
138 |
+
churn_by_IsActiveMember = df.groupby(['IsActiveMember', 'Exited']).size().unstack()
|
139 |
+
fig, ax = plt.subplots(figsize=(8, 4))
|
140 |
+
churn_by_IsActiveMember.plot(kind='bar', stacked=True, color=['pink', 'grey'], ax=ax)
|
141 |
+
plt.title('Customer Churn by IsActiveMember', fontsize=11)
|
142 |
+
plt.xticks(rotation=0)
|
143 |
+
plt.xlabel('IsActiveMember', fontsize=10)
|
144 |
+
plt.ylabel('Number of Customers', fontsize=10)
|
145 |
+
plt.legend(['Not Churn', 'Churn'],loc='lower center')
|
146 |
+
ax.set_xticklabels(['Inactive Member', 'Active Member'])
|
147 |
+
for container in ax.containers:
|
148 |
+
ax.bar_label(container, label_type='center')
|
149 |
+
st.pyplot(fig)
|
150 |
+
st.write('**Churn by Is Active Member or Not Statistics**')
|
151 |
+
st.write('Percentage of customers who churned with Inactive Member: 26.9 %')
|
152 |
+
st.write('Percentage of customers who churned with Active Member: 14.3 %')
|