🚀 Customer Churn Analysis & Prediction

Internship Project at SaiKet Systems

📊 Objective:
This project analyzes customer churn patterns, builds predictive models, and provides actionable insights to improve customer retention.

---

📌 Project Overview

✅ Performed Data Cleaning & Processing
✅ Exploratory Data Analysis (EDA) to Identify Churn Patterns
✅ Built Machine Learning Models for Churn Prediction
✅ Evaluated Model Performance Using Accuracy, Precision, Recall, and F1-Score
✅ Provided Business Recommendations to Reduce Churn

---

🛠 Technologies Used

• Google Colab ⚡
• Python 🐍
• Pandas, NumPy, Matplotlib & Seaborn 📊
• Scikit-Learn 🤖
• Power BI 📈 (For Data Visualization)

---

📈 Model Performance Comparison

Model	Accuracy	Precision	Recall	F1-Score
Logistic Regression	0.369	0.278	0.869	0.422
Decision Tree	0.656	0.429	0.912	0.584
Random Forest	0.735	0.000	0.000	0.000

---

🔍 Key Insights from Model Evaluation

• Logistic Regression has high recall (0.869), meaning it identifies churned customers well, but low precision.
• Decision Tree performs better overall, with higher accuracy (0.656) and balanced recall (0.912).
• Random Forest achieves the highest accuracy (0.735) but fails in precision and recall (0.0), possibly due to class imbalance.

---

💡 Next Steps & Improvements

🔹 Fix Class Imbalance: Use SMOTE (Synthetic Minority Over-sampling Technique) or re-sampling techniques.
🔹 Tune Hyperparameters: Use GridSearchCV for Decision Tree & Random Forest.
🔹 Try Other Models: Experiment with XGBoost, SVM, or Neural Networks for better performance.

---

📂 Project Files

📌 Main Notebook: PROJECT_UNDER_SAIKET_SYSTEM.ipynb

---