A Veerender, L Alekhya, G Shravani, M Sreekanth, Y Venkatesh 03 09 The HOTSTAR CUSTOMER CHURN ANALYSIS AND PREDICTION system is designed to analyze customer behavior and predict whether a customer is likely to stop using or subscribing to the Hotstar streaming platform With the rapid growth of online entertainment services customer retention has become an important challenge for streaming platforms Customers may discontinue their subscriptions due to factors such as subscription cost content availability viewing frequency service experience and competition from other platforms The proposed system uses historical customer data to identify patterns and factors associated with customer churn The collected data may include customer details subscription plan subscription duration viewing frequency watch time payment method content preferences and engagement level The data is first preprocessed by handling missing values duplicate records inconsistent data and categorical attributes Exploratory data analysis is performed to understand customer behavior and identify important churn patterns Machinelearning algorithms such as Logistic Regression Decision Tree Random Forest and Support Vector Machine can be used to build the prediction model The trained model predicts whether a customer is likely to Churn or Stay based on their characteristics and usage behavior Model performance can be evaluated using accuracy precision recall F1score and ROCAUC The system can also provide charts graphs and dashboards to visualize customer demographics subscription patterns engagement levels and churn rates The prediction results can help identify highrisk customers and support targeted customerretention strategies Overall the HOTSTAR CUSTOMER CHURN ANALYSIS AND PREDICTION system demonstrates how data analytics and machine learning can be used to understand customer behavior predict potential churn and support effective customer retention and datadriven business decisions Copyright (c) Journal of Science Engineering Technology and Management Science. All rights reserved 2026