PROJECT REPORT

ON
CHURN PREDICTION
Project SUBMITTED IN FULFILLMENT OF THE REQUIREMENT OF DEGREE OF
BACHELOR OF COMPUTER ENGINEERING(H)
OF

Faridabad,Haryana
2021-2025

SUBMITTED TO: SUBMITTED BY:

Dr.Umesh kumar Rahul Kumar(21001050014)

(ASSO.PROFESSOR)
 CERTIFICATE

It is certified that the project report titled “Churn Prediction ” being

submitted by Rahul Kumar student of BTECH(COMPUTER
ENGINEERING-hindi) (2021 to 2025) is genuine work carried
out by them under my supervision and guidelines.
This project work is the fulfillment of the requirement for the
degree of Bachelor of computer Engineering, J.C.Bose UST ,
Faridabad , Haryana.

Dr. Umesh kumar

Assistant Professor
 ACKNOWLEDGEMENT

Words are indeed inadequate to convey my deep sense of gratitude to all those who have
helped me in completing this project to the best of my ability. Being a part of this project
has certainly been a unique and a very productive experience on my part.

Amongst the wide panorama of people who provided me the inspiration, guidance and
encouragement, I take this opportunity to thank those who gave me indebted assistance
and constant encouragement for completing this project.

I would like to thank Dr.Umesh Kumar, JC BOSE UST,FARIDABAD for her continuous
help in completion of this project. She motivated me and was available whenever her
assistance was sought. She was actively involved throughout the project and was also
kind enough to tell me the strengths and weaknesses and how I could improve myself to
face the corporate world. Without her support the completion of this project would be
impossible.

RAHUL KUMAR
 TABLE OF CONTENTS

CANDIDATE'S DECLARATION
CERTIFICATE
ACKNOWLEDGEMENT

Introduction
This chapter gives an introduction about our motive behind making this
project and the benefits making this project and the benefits our project
offers to the users.

Project Descripition and methodology

This chapter gives an insight about technology and database used in this
Churn Prediction web Application.

Project Work Implementation

This chapter briefly explains all the functioning of project.In this,we have
shared screens available in the project and explains the usage of different
components an their implementation in the project.

CODE SNAPSOTS AND PAGE VIEW

This chapter briefly explains all the functioning of project.In this,we have
shared screens available in the project and explains the usage of different
components an their implementation in the project and we have shared the
screenshots of code.
1. INTRODUCTON
In today's competitive market landscape, retaining customers is as important as acquiring new ones.

Customer churn—defined as the loss of clients or subscribers— can significantly impact a

company's revenue and reputation. This is particularly true for industries such as

telecommunications, banking, subscription services, and e- commerce, where customer loyalty is

vital for sustained growth. Predicting churn helps businesses identify customers who are likely to

leave and allows them to take proactive measures to retain them. Effective churn prediction not only

mitigates revenue loss but also reduces the cost associated with acquiring new customers, which is

typically higher than retaining existing ones.

Customer churn can occur due to various factors, including poor service quality, lack of

engagement, competitive offerings, or dissatisfaction with the product. Understanding these factors

and predicting which customers are at risk of leaving is a strategic advantage. By analyzing

historical data and customer behavior, companies can detect early signs of churn and implement

targeted retention strategies, such as personalized offers or improved customer service. This project

aims to harness the power of data analytics and machine learning to create a robust churn prediction

model that provides actionable insights.

Churn prediction models often use machine learning algorithms that analyze patterns in customer

behavior, such as purchase frequency, engagement levels, and service interactions. These insights

help businesses anticipate churn and implement timely interventions, ensuring long-term customer

relationships and sustainable growth.

Whether in telecommunications, banking, e-commerce, or subscription services, churn prediction

has become a cornerstone of modern customer relationship management. For industries that thrive

on recurring revenue—such as telecommunications, e-commerce, subscription services, and

banking—minimizing churn is essential to sustaining growth and profitability

2.Project Description and Methodology

Python:-Python is used for server-side web development, software development, mathematics, and system

scripting, and is popular for Rapid Application Development and as a scripting or glue language to tie

existing components because of its high-level, built-in data structures, dynamic typing, and dynamic

binding. Program maintenance costs are reduced with Python due to the easily learned syntax and

emphasis on readability.

Jupyter Notebook:-Jupyter Notebook is an open-source, interactive computing environment that allows

users to create and share documents containing live code, visualizations, equations, and narrative text. It

supports multiple programming languages (like Python, R, and Julia) and is widely used for data analysis,

machine learning, scientific computing, and educational purposes. Jupyter's cell-based structure enables

incremental code execution, facilitating real-time feedback and collaboration.

Visual studio Code:-

Visual Studio is an integrated development environment (IDE) from Microsoft. It is used to

developcomputer programs including websites, web apps, web services and mobile apps. Visual Studio

uses Microsoft software development platforms such as Windows API, Windows Forms, Windows

Presentation Foundation, Windows Store and Microsoft Silverlight. It can produce both native code and

managed code.

Anaconda Navigator :-

Anaconda Navigator is a user-friendly desktop interface for managing packages, environments, and

applications within the Anaconda distribution. It allows users to easily install, update, and remove data

science libraries (like NumPy and TensorFlow) without using the command line.Navigator supports

environment creation and management, enabling isolated setups for different projects, which prevents

dependencyconflicts.
3.Project Work Implementation
Project Overview

This project focuses on predicting customer churn and retention using machine

learning techniques. The model leverages logistic regression to classify customers

as retained or churned. Key components include data preparation in Excel, model

building in Python, validation using k-fold cross- validation, and an interactive

Streamlit dashboard for result visualization.

Data Preparation in Excel

Data Collection:

Compiled historical customer data, including features like demographics, subscription

history, and usage patterns.

Data Cleaning:

Removed duplicates, handled missing values, and normalized data.

Feature Engineering:

Created relevant features such as customer tenure, average spending, and engagement

scores.

Data Export:

Saved the cleaned and processed data as a CSV file for further analysis in

Python.

Model Implementation using Logistic Regression

Environment Setup:

Libraries Used: pandas, numpy, sklearn, streamlit, matplotlib, seaborn

Data Import: Loaded the CSV file into a pandas DataFrame.

Logistic Regression Steps:

Data Splitting:

Split data into training (70%) and testing (30%) sets.

Model Training:

Implemented logistic regression using scikit-learn’s LogisticRegression class.

Hyperparameter Tuning:

Adjusted parameters like C (regularization strength) and penalty terms to optimize

performance.

Model Evaluation Metrics:

Accuracy, precision, recall, F1-score, and confusion matrix.

K-Fold Cross-Validation

Objective:

Evaluate model stability and performance.

Implementation:

Used k-fold cross-validation (with k=5) to divide the data into 5 subsets. The

model was trained and tested 5 times, each time using a different subset for

testing and the remaining for training.

Mean Accuracy: Calculated the average accuracy across all folds to assess model

reliability.

Deployment with Streamlit

Purpose:

Create a user-friendly web application to visualize predictions and insights.

Streamlit Workflow:

UI Elements:

Uploaded data via file uploader.

Dropdowns and sliders for user inputs (optional).

Prediction Display:

Provided customer retention probability and classification.

Displayed performance metrics and visualizations (e.g., ROC curve, confusion matrix).
4.Snapshots of user interface:
5.Snapshots of code:
6. REFRENCES

Following sources have been sought for the preparation of this

report:
1. CHATGPT
2. DOCUMENTATION
3. TRAIN MODEL INSTRUCTION AND DOCUMNETATION ON JUPYTER NOTEBOOK
4. https://chat.openai.com/c/4502e748-f10f-4582-9d82- 1a97817b5fd5