Mathew K AnalyticsLesson 73 · Data Science Projects
Customer Churn Prediction: Step-by-Step Classification Models in Python
Welcome to this hands-on notebook focused on predicting customer churn. You will learn how to use Python and machine learning to find out which customers…
- CourseData Science Projects
- Lesson73 of 33
- Video17 min
- FormatJupyter notebook · 17 code cells
What you'll learn
Data
No separate download needed — the notebook creates or downloads everything it uses.
📓 Full notebook
Download .ipynbWeek 45: Customer Churn Prediction with Classification Models#
Welcome to this hands-on notebook focused on predicting customer churn.
You will learn how to use Python and machine learning to find out which customers are likely to leave a company.
We will use the Telco Customer Churn Dataset, explore the data, clean it, build classification models, and interpret results.
Customer churn prediction helps businesses keep their customers by taking action before someone leaves.
Let us get started!
# Suppress warnings for a clean notebook
import warnings; warnings.filterwarnings("ignore")
import numpy as np
np.random.seed(42)
# Data setup (Telco Customer Churn Dataset)
import pandas as pd
url = 'https://raw.githubusercontent.com/IBM/telco-customer-churn-on-icp4d/master/data/Telco-Customer-Churn.csv'
df = pd.read_csv(url)
print(df.shape)
print(df.head(3))
Step 1: Understanding the Data#
Let us take a closer look at the churn data.
- Each row represents a customer.
- Columns describe customer demographics, contract information, services used, and whether they have churned.
The 'Churn' column says if a customer has left or stayed.
# Check if there are missing values
missing = df.isnull().sum()
print(missing[missing > 0])
# View summary statistics for numerical columns
print(df.describe())
# View column types and unique values for 'Churn'
print(df.dtypes)
print(df['Churn'].unique())
# How many customers have churned?
print(df['Churn'].value_counts())
Step 2: Data Cleaning and Preprocessing#
The data sometimes needs cleaning before it can be used in machine learning.
- Fix or remove empty values
- Convert text data to numbers
- Pick the columns we need
Let us start the cleaning process.
# Convert 'TotalCharges' to numeric and handle errors
df['TotalCharges'] = pd.to_numeric(df['TotalCharges'], errors='coerce')
# Drop rows with missing 'TotalCharges'
df = df.dropna(subset=['TotalCharges'])
# Encode the 'Churn' column to 1 for 'Yes' and 0 for 'No'
df['Churn'] = df['Churn'].map({'Yes':1, 'No':0})
# Select columns to use for modeling (drop ID columns)
X = df.drop(['customerID', 'Churn'], axis=1)
y = df['Churn']
# Convert all category columns to dummy variables
X = pd.get_dummies(X)
Step 3: Split Data for Training and Testing#
To check if our model really works, we split our data so it learns from some (training) and is tested on the rest (testing).
# Split into 70% train and 30% test, random_state=42 for repeatability
from sklearn.model_selection import train_test_split
X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.3, random_state=42)
Step 4: Train a Logistic Regression Model#
Let us start with a simple and powerful classification method called Logistic Regression.
- Logistic Regression works well for yes/no problems.
from sklearn.linear_model import LogisticRegression
# Create and train the model
clf = LogisticRegression(max_iter=1000)
clf.fit(X_train, y_train)
# Predict churn using our trained model
y_pred = clf.predict(X_test)
# Evaluate accuracy
from sklearn.metrics import accuracy_score
acc = accuracy_score(y_test, y_pred)
print(f"Accuracy: {acc:.2f}")
# Print confusion matrix for deeper insight
from sklearn.metrics import confusion_matrix
cm = confusion_matrix(y_test, y_pred)
print(cm)
# Show precision and recall for balanced understanding
from sklearn.metrics import classification_report
print(classification_report(y_test, y_pred, target_names=['Stayed','Churned']))
# Try a Decision Tree for comparison
from sklearn.tree import DecisionTreeClassifier
tree = DecisionTreeClassifier(random_state=42, max_depth=4)
tree.fit(X_train, y_train)
y_pred_tree = tree.predict(X_test)
print('Decision Tree accuracy:', accuracy_score(y_test, y_pred_tree))
Quick Recap and Your Next Steps#
In this notebook you learned to:
- Load a real churn dataset
- Clean and prepare your data
- Build and test both logistic regression and decision tree models
- Measure accuracy and see where the models work best
To practice, try changing parameters, use more models, or visualize top features.
If you learned something valuable, subscribe to the channel and keep exploring data mining!
Found this useful?
All lessons, notebooks and datasets here are free. If they helped you, a coffee keeps new lessons coming.
Related lessons
23 minCustomer Churn Prediction: Step-by-Step Python Tutorial Using Real Data
14 minMaster Accuracy, Precision & Recall for Evaluating Classification Models in Python
12 minUnderstanding ROC Curves and AUC for Evaluating Classification Models in Python
13 min