Customer Churn is a major concern for businesses, leading to potential revenue loss. This study develops and compares data driven models using a comprehensive dataset to predict churn. Through graphs and in-depth visualizations, the study analyses key features that contribute to the churn. It addresses data imbalance with oversampling techniques and highlights the importance of feature selection in determining the most relevant features. Among the tested algorithms, Random Forest Classifier emerged as most suitable giving an ROC-AUC score of 0.99. Methods like feature scaling and hyperparameter tuning are employed to optimize the performance of the model.